PALACKÝ UNIVERZITY OLOMOUC FAKULTY OF SCIENCE DEPARTMENT OF BOTANY

Anna Heinken-Šmídová

Study of factors influencing population dynamics of the plant species Ligularia sibirica (L.) Cass.

PhD. THESIS

Supervisor: Doc. RNDr. Zuzana Münzbergová, Ph.D.

Olomouc 2012

Bibliografická identifikace

Jméno a příjmení autora: Anna Heinken-Šmídová Název práce: Studie faktorů ovlivňující dynamiku druhu Ligularia sibirica (L.) Cass. Typ práce: doktorská Pracoviště: Katedra botaniky – UP Olomouc; Botanický ústav akademie věd ČR Studijní program: Biologie Studijní obor: Botanika Školitel: Doc. RNDr. Zuzana Münzbergová, Ph.D. Rok obhajoby: 2012 Abstrakt: Dobrá znalost populační biologie, genetické variability a stanovištních charakteristik je nezbytná pro efektivní ochranu vzácných druhů rostlin. Pro svou studii jsem si zvolila kriticky ohrožený druh chráněný směrnicí o stanovištích (příloha II), Ligularia sibirica (L.) Cass.. Hlavním cílem studie bylo stanovit genetickou variabilitu, hlavní stanovištní charakteristiky lokalit studovaného druhu a zhodnotit jejich význam spolu s velikostí populací pro populační dynamiku L. sibirica v České republice a na Slovensku. Populační dynamika byla studována pomocí populačních přechodových matic po dobu čtyř let. Většina sledovaných populací prospívá dobře a pouze populace rostoucí na odvodněných stanovištích mají klesající trend. Analýzy elasticity ukázaly, že přechody, které nejvíce přispívají k růstové rychlosti populací, se liší dle zachovalosti stanoviště. Genetická variabilita byla stanovena pomocí metody isoenzymů. Výsledky ukázaly, že populace L. sibirica jsou schopny uchovat dostatek genetické variability a že se většina této variability nachází uvnitř populací. Nalezena byla také pozitivní korelace mezi genetickou variabilitou a velikostí populací druhu. Výsledky dále ukázaly, že L. sibirica má schopnost růst v široké škále na živiny různě bohatých stanovištích. Populace druhu rostoucí na stanovištích chudých na živiny prosperují lépe než populace rostoucí na stanovištích bohatých na živiny, a to i přesto, že je velikost jedinců menší a mají nižší produkci semen. Nakonec jsem regresní analýzou zjistila, který ze studovaných faktorů (genetická variabilita, stanovištní podmínky nebo velikost populace) ovlivňuje vitalitu populací L. sibirica. Ukázalo se, že tímto faktorem jsou stanovištní podmínky. Výsledky ukazují, že malá velikost populace a nižší genetická variabilita prozatím nejsou hlavní problém při ochraně druhu Ligularia sibirica. Ochrana tohoto vzácného druhu by se měla zaměřit především zachování vhodných stanovištních podmínek a na obnovení stanovištních podmínek odvodněných lokalit.

Klíčová slova: Ligularia sibirica; vzácný rostlinný druh; životaschopnost populací; genetická variabilita; velikost populací; kvalita stanoviště Počet stran: 133 Jazyk: anglický

3

Bibliographic identification

Author’s first name and surname: Anna Heinken-Šmídová Title of the Thesis: Study of factors influencing population dynamics of the plant species Ligularia sibirica (L.) Cass. Type of Thesis: Ph.D. Thesis Department: Department of Botany – UP Olomouc; Institute of Botany of the ASCR Study program: Biology Field of Study: Botany Supervisor: Doc. RNDr. Zuzana Münzbergová, Ph.D. Year of defence: 2012 Abstract: Good knowledge of population biology, genetic diversity, and environmental factors is necessary for efficient conservation of rare species. For my study I chose Ligularia sibirica (L.) Cass., a critically endangered plant species protected under the EU Habitats Directive (Annex II). The main aim of the study was to estimate the genetic diversity and the main habitat conditions of the localities, and assess their importance together with population size for the population dynamics of L. sibirica in the Czech and Slovak Republic. The population dynamics was studied for four years using population transition matrices. Most of the studied populations of L. sibirica were performing well and only those growing in deteriorating habitats due to drainage showed a decreasing trend. Analyses of elasticity showed that transitions that most contribute to the population growth rate vary with habitat conditions. The genetic diversity was estimated using allozyme electrophoresis. The findings suggest that populations can retain sufficient amount of genetic variability and that most of the variability is within populations. However, the results also revealed a significant positive correlation between genetic diversity and population size. Further, the results showed that L. sibirica has the ability to grow and compete under various nutrient regimes. Populations growing in nutrient-poor habitats were performing better than populations growing in nutrient-rich habitats, even though the individuals are smaller and have lower seed production in nutrient-poor conditions. Comparing the importance of habitat conditions, population size and genetic diversity for the performance of the L. sibirica population, the habitat conditions were identified as the most important factor. The results indicate that small population sizes and low/reduced genetic diversity are not yet a major problem for Ligularia sibirica and protection of this species should focus on preserving habitat conditions of nutrient-poor sites and restoring habitat conditions of drained localities.

Key-words: Ligularia sibirica; rare plant species; population viability; genetic diversity; population size; habitat quality Number of pages: 133 Language: English

5

Declaration

I hereby declare that I made this thesis independently, using the listed references, or in the co- operation with other authors of the papers. I did submit neither the thesis nor any part of it to acquire any other academic title.

Anna Heinken-Šmídová

Contributions of the co-authors

Chapter I: Heinken-Šmídová A. & Münzbergová Z.: Population dynamics of the endangered, long-lived perennial species, Ligularia sibirica. – accepted to Folia Geobotanica

A. Heinken-Šmídová collected the field data, performed their analyses and wrote the chapter. Z. Münzbergová gave advices on methodology and statistical analyses, reviewed the chapter and her comments helped to improve the manuscript.

Chapter II: Šmídová A., Münzbergová Z. & Plačková I. (2011): Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae). – Flora 206: 151-157.

A. Heinken-Šmídová collected the field data, participated on the laboratory work, performed the analyses and wrote the chapter. Z. Münzbergová gave advices on methodology and statistical analyses, reviewed the chapter and her comments helped to improve the manuscript. I. Plačková conducted the laboratory allozyme analyses and gave advices for the evaluation of the results.

Chapter III: Heinken-Šmídová A. & Münzbergová Z.: The habitat requirements of the endangered wetland species Ligularia sibirica. – Manuscript.

A. Heinken-Šmídová collected the field data, performed their analyses and wrote the chapter. Z. Münzbergová gave advices on methodology and statistical analyses, reviewed the chapter and her comments helped to improve the manuscript. The soil and leaf biomass analyses were performed in the soil laboratory of the Institute of Botany of the ASCR.

Chapter IV: Heinken-Šmídová A. & Münzbergová Z.: Habitat conditions are better determinants of population performance of the perennial herb Ligularia sibirica than population size and genetic diversity. – Manuscript.

A. Heinken-Šmídová performed the statistical analyses and wrote the chapter. Z. Münzbergová gave advices on statistical analyses, reviewed the chapter and her comments helped to improve the manuscript.

Given the co-authorship of Zuzana Münzbergová who reviewed all chapters and her comments helped to improve the manuscripts, I will use the first person plural in the following.

All authors are aware of the above text. They agree with the wording of the statement.

7

Acknowledgements

I would like to thank to my supervisor, Zuzana Münzbergová for her help, advices and securing the funding of this study, and to Vlastík Rybka for inspiring me to study Ligularia sibirica. I have to thank to A. Leskovjanská from the National Park Slovensky raj, E. Gojdičová from the State Nature Conservancy of the Slovak Republic, Prešov and P. Turis from the National Park Nizke Tatry who provided me with all information about the Slovak localities and helped me very much with my field work in the Slovak Republic. I also thank the workers in the protected areas of the Czech Republic in which Ligularia sibirica occurs for providing me with information about the localities. My thanks also go to the members, especially to Ivana Plačková, of the isozyme laboratory in the Institute of Botany, Academy of Science of the Czech Republic in Průhonice for introducing me into the allozyme analyses. My thanks belong to my family for supporting me in various ways. Last but not least I would like thank my husband Thilo for his great help with the field work, scientific advices and ideas to my work and that he stood by me all the time and encouraged me to continue in the work. Finally I must thank my daughter Klara, born 24th November 2011, for giving me a few moments to finish the last lines of the thesis.

This thesis was supported by the research project from the Czech Ministry of Education 2B06178 and partly also 0021620828 and by the research project from the Grant Agency of Academy of Science AV0Z60050516.

Due to the conservation status of Ligularia sibirica, the author obtained research permission from the Ministries of the Environment of the Czech and Slovak Republic to enter the localities of populations of this species, and for scientific research on these localities.

9

Table of contents

GENERAL INTRODUCTION AND OUTLINE OF THE THESIS 13

CHAPTER I Population dynamics of the endangered, long-lived perennial species, Ligularia sibirica 21 Heinken-Šmídová A. & Münzbergová Z. Folia Geobotanica [accepted]

CHAPTER II Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae) 55 Šmídová A., Münzbergová Z. & Plačková I. (2011) Flora 206: 151-157

CHAPTER III The habitat requirements of the endangered wetland species Ligularia sibirica 73 Heinken-Šmídová A. & Münzbergová Z. [manuscript]

CHAPTER IV Habitat conditions are better determinants of population performance of the perennial herb Ligularia sibirica than population size and genetic diversity 99 Heinken-Šmídová A. & Münzbergová Z. [manuscript]

CONCLUSIONS 127

CURRICULUM VITAE 131

11

GENERAL INTRODUCTION

AND OUTLINE OF THE THESIS

General introduction and outline

General introduction

In Europe currently one thousand plant species are considered as endangered. However, only a few hundreds of these species have a legal protection (Ozinga & Schaminée 2005), which is mainly on the national level. Since 1992, there is a list of the European Union for the protection of rare plant species known as the Habitats Directive, Annex II (Council of the European Community 1992). Most of the species listed here are endemics or relicts, whose distribution in Europe is either limited to a small area, or they require specific habitat conditions. The main aim of the protection of above mentioned rare plant species in terms of their conservation and management is to maintain viable populations (Soulé 1987). One of the most frequently used methods for assessing the viability of populations is Population Viability Analyses, PVA (Menges 2000). PVA is mainly concerned with population growth rate and with the risk of extinction of populations. In addition, PVA can also be used to identify critical life stages of the populations (Ramula 2006). A large number of factors can contribute to the endangerment or even to the extinction of entire populations of species or species themselves a large number of factors can contribute. Some of these factors may be specific for a particular plant species (e.g. specialized pollination or germination). However, based on the results of several studies on rare or declining plant species, we can identify groups of factors that affect the viability of populations of most species (e.g. Fischer & Matthies 1998; Oostermeijer et al. 1998; Vergeer et al. 2003; Lienert 2004). They are the quality of habitat, population size and genetic diversity (see Fig. 1). These factors may act either individually or together and their decrease may lead to an ‘extinction vortex’ for a plant species (Ellstrand & Elam 1993).

Fig.1: Theoretical relationships between habitat quality, population size, genetic diversity and population viability in plant species (adapted from de Vere et al. 2009). Bold face: factors studied in this thesis.

14 General introduction and outline

Changes in habitat quality are often said to be the most important factors affecting the viability of populations (Brys et al. 2005; Schleuning & Matthies 2009). These changes are linked with changes in the availability of nutrients and water (Colling et al. 2002; Brys et al. 2005), succession or a change in management (Oostermeijer et al. 1996). Negative changes in habitat quality often lead to lower viability of the populations, which in turn leads to a reduction in population size (Ouborg et al. 2006). Changes in population size may be due to either reduced fitness or due to destruction of the population (Oostermeijer et al. 2003; Leimu et al. 2006). These changes are often associated with changes of habitat quality. On the other hand, reduced population size may lead to the disruption of interactions between plants and pollinators (Lienert 2004; Aguilar et al. 2006) and a loss of genetic diversity (e.g. Frankham 2005; Oostermeijer et al. 2003; Lienert 2004; Leimu et al. 2006). Small populations of animal-pollinated plants can be less attractive to pollinators than large populations and consequently lower pollinator visitation can lead to lower seed production (Ågren 1996; Lienert 2004; Aguilar et al. 2006). The loss of genetic diversity has two main effects: genetic drift and increased inbreeding. Under genetic drift, the frequencies of rare alleles within populations are affected and they can be lost from the gene pool (Amos & Balmford 2001). In small populations exists a higher probability of inbreeding, when genetically related individuals mate (Keller &Waller 2002) which may lead to inbreeding depression (e.g. Young et al. 1999; Kéry et al. 2000). In the short term reduction of genetic diversity as in the case of reduced population size, is associated with lower fitness (e.g. low production and germination of seeds, high mortality and poor growth of seedlings) (Ellstrand & Elam 1993; Oostermeijer et al. 2003). In long term, low genetic diversity may reduce the ability of species to adapt to changing habitat conditions and consequently lead to the extinction of the entire population (Ouborg et al. 2006).

In the past decades many studies on rare and recently declining plant species have been conducted which looked at the correlations between factors such as habitat conditions, population size, genetic diversity and single vital rates, especially seed set, flowering probability, and germination (e.g. Fischer & Matthies 1998; Oostermeijer et al. 1998; Brys et al. 2004; Hensen et al. 2005). Also, a number of studies on rare species focused on population dynamics (e.g. Brys et al. 2005; Colling & Matthies 2006; Schleuning & Matthies 2009). However, very few studies considered all those factors such as habitat conditions, genetic diversity, population size and their possible interaction together (but see Vergeer et al. 2003; de Vere et al. 2009) even though it was shown that by knowing the effect of a factor on single vital rates, we cannot easily predict its effect on overall population dynamics (Ehrlén 2003; Münzbergová 2006). In my thesis I carried out a detailed study on the population biology of a relict plant species, Ligularia sibirica (L.) Cass., in the Czech and Slovak Republic. The aim of my study was to find

15 General introduction and outline out which of the factors that are considered to have the greatest effect on the viability of plant species populations (i.e. habitat quality, population size and genetic diversity) influence single parts of the life cycle or population viability of L. sibirica. To answer this question I studied the full life cycle of L. sibirica in order to gain information about population growth rate, to identify critical life stages and to estimate the risk of extinction of populations in relation to habitat conditions, population size and genetic diversity. Ligularia sibirica belongs between co-called “old rare species” (species that are naturally rare in a specific area, occurring in small, isolated populations (sensu Huenneke 1991)) and is protected under the EU Habitats Directive (Annex II). Such species are until now understudied compared to recently declining plant species (but see e.g. Paschke et al. 2002; Picó & Riba 2002; Garcia 2003; Pino et al. 2007). The reason might be of practical nature, as rare and narrowly distributed species have a limited number of occurrences and/or individuals which might hinder obtaining sufficient sample sizes. Ligularia sibirica is a long-lived perennial herb of the Asteraceae family with mixed-mating breeding system. It occurs in a variety of wetlands, such as humid grasslands, calcareous fens, transitional mires and alluvial woodlands with the groundwater table at the surface (Slavík 2004). The species has a wide Eurosiberian distribution range. Its main continuous distribution is from East Asia to Southern Siberia and to the European part of Russia. In the remaining part of Europe, there are few disjunct, isolated populations in Estonia, Latvia, Poland, Hungary, Romania, Croatia, Bulgaria, the Slovak Republic, the Czech Republic, Austria, and France (Meusel & Jäger 1992).

References: Ågren, J. (1996): Population size, pollinator limitation, and seed set in the self-incompatible herb Lythrum salicaria. Ecol., 77: 1779-1790. Aguilar, R., Ashworth, L., Galett, L. & Aizen, M.A. (2006): Plant reproductive susceptibility to habitat fragmentation: review and synthesis through a meta-analysis. Ecol. Lett., 9: 968-980. Amos, W. & Balmford, A. (2001): When does conservation genetics matter? Heredity, 87: 257- 265. Brys, R., Jaquemyn, H., Endels, P., de Blust, G. & Hermy, M. (2005): Effect of habitat deterioration on population dynamics and extinction risks in a previously common perennial. Conserv. Biol., 19: 1633-1643. Brys, R., Jaquemyn, H., Endels, P., van Rossum, F., Hermy, M., Triest, L., de Bruyn, L. & de Blust, G. (2004): Reduced reproductive success in small populations of the self-incompatible Primula vulgaris. J. Ecol., 92: 5-14. Colling, G. & Matthies, D. (2006): Effects of habitat deterioration on population dynamics and extinction risk of an endangered, long-lived perennial herb (Scorzonera humilis). J. Ecol., 94: 959–972. Colling, G., Matthies, D. & Reckinger, C. (2002): Population structure and establishment of the threatened long-lived perennial Scorzonera humilis in relation to environment. J. Appl. Ecol., 39: 310-320. Council of European Communities (1992): Council Directive 92/43/EEC of 21 May on the conservation of natural habitats and of wild fauna and flora. Off. J. Eur. Communities, 35: 7- 50.

16 General introduction and outline

De Vere, N., Jongejans, E., Plowman, A. & Williams, E. (2009): Population size and habitat quality affect genetic diversity and fitness in the clonal herb Cirsium dissectum. Oecologia, 159: 59-68. Ehrlén, J. (2003): Fitness components versus total demographic effects: evaluating herbivore impacts on a perennial herb. Am. Nat., 162: 796-810. Ellstrand, N. C. & Elam, D. C. (1993): Population genetic consequences of small population size: implication for plant conservation. Annu. Rev. Ecol. System., 24: 217-242. Fischer, M. & Matthies, D. (1998): Effects of population size on performance in the rare plant Gentianella germanica. J. Ecol., 86: 195-204. Frankham, R. (2005): Genetics and extinction. Biol. Conserv., 126: 131-140. García, M.B. (2003): Demographic viability of a relict population of the critically endangered plant Borderea chouardii. Conserv. Biol., 17: 1672-1680. Hensen, I., Oberprieler, C. & Wesche, K. (2005): Genetic structure, population size, and seed production of Pulsatilla vulgaris Mill. (Ranunculaceae) in central Germany. Flora, 200: 3– 14. Huenneke, L.F. (1991): Ecological implications of genetic variation in plant populations. In: Falk, D.A. & Holsinger, K.E. (eds) Genetics and conservation of rare plants. Oxford University Press, Oxford, pp 31-44. Keller, L.F. & Waller, D.M. (2002): Inbreeding effects in wild populations. Trends Ecol. Evol., 17: 230-241. Kéry, M., Matthies, D. & Spillmann, H.-H. (2000): Reduced fecundity and offspring performance in small populations of the declining grassland plants Primula veris and Gentiana lutea. J. Ecol., 88: 17-30. Leimu, R., Mutikainen, P., Koricheva, J. & Fischer, M. (2006): How general are positive relationships between plant population size, fitness and genetic variation? J. Ecol., 94: 942- 952. Lienert, J. (2004): Habitat fragmentation effects on fitness of plant populations – a review. J. Nat. Conserv., 12: 53–72. Menges, E.S. (2000): Population viability analyses in plants: challenges and opportunities. Trends Ecol. Evol., 15: 51-56. Meusel, H. & Jäger J. (1992a, 1992b.): Vergleichende Chorologie der zentraleuropäischen Flora. Vol. 3. (1992a) Text, (1992b) Karten. G. Fischer, Jena. 491p. Münzbergová, Z. (2006): Effect of population size on prospect of species survival. Folia Geobot., 41: 137–150. Oostermeijer, J.G.B., Brugman, M.L., de Boer, E.R. & den Nijs, H.C.M. (1996): Temporal and spatial variation in the demography of Gentiana pneumonanthe, a rare perennial herb. J. Ecol., 84: 153-166. Oostermeijer, J.G.B., Luijten, S.H. & den Nijs, J.C.M. (2003): Integrating demographic and genetic approaches in plant conservation. Biol. Conser. 113: 389-398. Oostermeijer, J.G.B., Luijten, S.H., Křenová, Z.V. & den Nijs, J.C.M. (1998): Relationship between population and habitat characteristics and reproduction of the rare Gentiana pneumonanthe L.. Conser. Biol., 12: 1042-1053. Ouborg, N.J., Vergeer, P. & Mix, C. (2006): The rough edges of the conservation genetics paradigm for plants. J. Ecol., 94: 1233-1248. Ozinga, W.A. & Schaminée, J.H.J. (eds.) (2005): Target species – Species of European concern. A database driven selection of plant and animal species for the implementation of the Pan European Ecological Network. Wageningen, Alterra, Alterra-report 1119. 193p. Paschke, M., Abs, C. & Schmid, B. (2002): Relationship between population size, allozyme variation, and plant performance in the narrow endemic Cochleria bavarica. Conserv. Genet., 3: 131-144. Picó, F.X. & Riba, M. (2002): Regional-scale demography of Ramonda myconi: Remnant population dynamics in a preglacial relict species. Plant Ecol., 161: 1–13. Pino, J., Picó, F.X. & de Roa, E. (2007): Population dynamics of the rare plant Kosteletzkya pentacarpos (Malvaceae): a nine-year study. Bot. J. Linn. Soc., 153: 455–462. Ramula, S. (2006): Population viability analysis for plants: Practical recomendations and applications. Doctoral dissertation, Jannes Kuvertproffset HB, Stockholm, Sweden.

17 General introduction and outline

Schleuning, M. & Matthies, D. (2009): Habitat change and plant demography: Assessing the extinction risk of a formerly common grassland perennial. Conser. Biol., 23: 174-183. Slavík, B. (2004): Ligularia Cass. In Slavík, B. & Štěpánková, J. (eds) Květena České republiky, Vol. 7. Academia, Praha, pp 306–309. Soulé, M.E. (1987): Viable populations for conservation. Cambridge University Press, Cambridge. Vergeer, P.V., Rengelink, R., Copal, A. & Ouborg, N.J. (2003): The interacting effects of genetic variation, habitat quality and population size in performance of Succisa pratensis. J. Ecol., 91: 18-26. Young, A., Boyle, T. & Brown, T. (1999): The population genetic consequences of habitat fragmentation for plants. Trends Ecol. Evol., 11: 413-418.

18 General introduction and outline

Outline of the contents of the thesis

This thesis consists of four chapters. These represent results aimed on the detailed study of a relict plant species protected under the EU Habitats Directive (Annex II), Ligularia sibirica (L.) Cass.. This species is a long-lived perennial herb inhabiting wetlands with mixed-mating breeding system, widely distributed from East Asia to European Russia, with few isolated relict populations in the remaining part of Europe.

The general objectives of this study can be summarized as follows:

Chapter I: “Population dynamics of the endangered, long-lived perennial species, Ligularia sibirica“

The objective of this chapter was to estimate differences in performance among a wide range of populations of Ligularia sibirica differing mainly in their size and nutrient availability of their habitats. Additionally, we determined the population growth rate, longevity and risk of extinction of each population and identified the specific vital rates that most affect population growth rate.

Chapter II: “Genetic diversity of a relict plant species, Ligularia sibirica (L.)“

The aim of this chapter was to examine genetic diversity of 20 populations of our study species, Ligularia sibirica, in the Czech and Slovak Republic, in relation to population size and geographic distance. Specifically, we aimed to interprete the results in relation to the biology of L. sibirica and discuss how the present pattern of genetic diversity is affected by the possible history of the populations.

Chapter III: “The habitat requirements of the endangered wetland species Ligularia sibirica“

In this chapter we studied the habitat requirements of Ligularia sibirica. We described plant species composition, investigated soil and leaf biomass chemistry at all its known sites in the Czech and Slovak Republic. We compared habitat conditions with those of populations in other European countries and asked how fitness and population size of L. sibirica are affected by the environmental variables.

19 General introduction and outline

Chapter IV: “Habitat conditions are better determinants of population performance of the perennial herb Ligularia sibirica than population size and genetic diversity“

The main purpose of this chapter was to investigate how habitat conditions, population size and genetic diversity influence the population performance of Ligularia sibirica. In this study we combined data gained in the previous studies (i.e. Chapter I, Chapter II and Chapter III). By joining data on the full life cycle of L. sibirica with data on habitat conditions, genetic diversity and population size, we asked what is the effect of year, habitat conditions, genetic diversity and population size on single vital rates and on the whole life cycle of the species. Because the effects of habitat conditions, genetic diversity and population size can be partly correlated, we tested both the overall effects (i.e. the effect of each of the three factors separately) and the pure effects of each of the factors using the other two as covariates.

20

CHAPTER I

Population dynamics of the endangered, long-lived perennial species, Ligularia sibirica

Heinken-Šmídová A. & Münzbergová Z. Folia Geobotanica [accepted]

Chapter I

Abstract Ligularia sibirica is a relict wetland perennial plant species of the Czech and Slovak Republic. Explaining variation in population growth rate and identifying the causes of that variation is important for effective protection of such an endangered species. Matrix models based on four years of data of 11 populations were used to identify the pattern of variation in the demographic vital rates of this species, and to examine the causes of the variation such as population size and habitat type. Further, the matrix model was used to determine the population growth rate, longevity and risk of extinction of each population and to identify the specific vital rates that most affect population growth rate. The results showed that population growth rates were significantly different between years and populations. Temporal variation was mostly due to variable survival of adult individuals, while spatial variation was mainly driven by fertility of one small currently expanding population. Further, most studied populations of L. sibirica are performing well and only those growing in nitrogen-rich habitats have a high extinction risk. The results also indicate that all populations have low adult mortality, long-lived individuals (61.3 years on average) and some populations also show features of remnant populations (i.e. the persistence of populations in severe conditions in spite of no reproduction). Our results imply that detailed demographic data are needed to understand the long-term prospects of these populations. These data may serve as an early warning system for this species long before an obvious decline occurs in the populations.

Keywords: Longevity; LTRE; Matrix population model; Nutrient enrichment; Population growth rate; Population size; Population viability analysis

Introduction One task of conservation biology is to develop practical approaches to prevent extinction of endangered species (Primack 2002). To do this, all biological information relevant to the survival and recovery of the endangered species must be gathered (Brussard 1991; Schemske et al. 1994) and summarized using population viability analyses (PVA). PVA are used to assess extinction risks, project growth of populations, and identify which life stages are the most important for population growth (Menges 2000; Reed et al. 2002). Many PVA have been conducted on plants over the past few decades (reviewed in Menges 2000) including studies on various endangered plant species (Oostermeijer et al. 2003). Despite the high number of existing PVA of various endangered species, we still know relatively little about differences in performance of populations under various habitat conditions, management practices and in populations of various sizes. This is because most studies focus only on one or very few populations (but see Fréville et al. 2004). Understanding how population

22 Chapter I dynamics vary in space depending on local habitat conditions and size is key for understanding species habitat requirements and setting proper conservation targets. Changes in population size and in habitat conditions such as nutrient enrichment and management practices were previously reported to have an impact on single vital rates. Specifically, population size, nutrient availability and management practices may effect seed production (Oostermeijer et al. 1998; Lennartsson & Oostermeijer 2001; Münzbergová 2006), germination rate (Colling et al. 2002; Lindborg & Ehrlén 2002; Vergeer et al. 2003) and survival of seedlings (Colling et al. 2002). The same factors are reported to have an effect on the stasis of individuals (Hegland et al. 2001; Tomimatsu & Masashi 2010) and on growth of individuals (Dahlgren & Ehrlén 2009). This study aimed to estimate the differences in performance among a range of populations differing mainly in their size and nutrient availability at their habitats over a wide range of populations of a wetland long-lived perennial herb, Ligularia sibirica (L.) Cass.. Ligularia sibirica is a relict species protected by the Habitats Directive, Annex II of the Council of the European Community (1992), i.e. it is among the species protected throughout the European Union. In the study countries, the Czech and Slovak Republic, populations range from 123 to 50,000 adult individuals, and are of different habitat type, i.e. nitrogen enrichment levels (Hendrych 2003; Šmídová 2006, 2009). We studied the population dynamics in 11 populations of this species over four years and wanted to answer the following questions: i) What is the spatial and temporal variation between different populations of L. sibirica populations? ii) Can it be explained by population size and habitat type? iii) Which vital rates contribute the most to the observed variation in population growth rate? Because field observations in our system indicated a high degree of herbivory in our populations, and many previous studies have demonstrated that different type of herbivores can have a dramatic effect on the performance of plant populations (e.g. Crawley 1989; Münzbergová 2005) we included the effects of herbivory into our demographic models.

Material and methods

Study species Ligularia sibirica (L.) Cass. (Asteraceae) is a perennial hemicryptophyte 124.25±43.00 cm high (mean ± s.d.) with ground rosettes of leaves and a short rhizome. During flowering from mid- July to the end of August, L. sibirica creates 1.91±1.73 flowering stems (mean ± s.d.) with 18.79±8.42 flower heads per inflorescence (mean ± s.d.). The flower heads contain 32.13±7.26 achenes (mean ± s.d.) (unpubl. data). The achenes (further referred to as seeds) are of oval shape and weigh 2.16 mg±0.80 (mean ± s.d.) (Šmídová et al. 2011). Their pappus forms a tuft of hair longer than the achenes (Hegi 1929; Slavík 2004). The seeds are dispersed over a short distance by

23 Chapter I either wind or gravity. The terminal velocity of L. sibirica seeds varies between 0.86 and 1.18 m s-1 (Šmídová et al. 2011). There is a pre-dispersal seed predator from the family Gelechiidae (Lepidoptera) with larvae that develop inside the flower head and internally consume the seeds of L. sibirica (Elsner & Patočka, pers. comm.). Ligularia sibirica has hermaphrodite entomogamous flower heads (Slavík 2004). Sexual reproduction is highly predominant. Clonal growth of the ‘phalanx’ type, i.e. in the form of a tightly-packed advancing front of ramets with a slow radial spread (sensu Doust 1981), may also occur. Ligularia sibirica is a plant species that prefers full sunlight, but it also grows under open tree canopy where seed production is lower (Kukk 2003). It occurs in a variety of wetlands, such as humid grasslands, calcareous fens, transitional mires and alluvial woodlands with the groundwater table at the surface (Slavík 2004). Ligularia sibirica is regarded as a postglacial relict (Hendrych 2003; Šmídová et al. 2011), classified as a ‘critically endangered’ species in the Czech Republic (Procházka 2001) and as a ‘vulnerable’ species in the Slovak Republic (Feráková et al. 2001). We obtained permission from the Ministries of the Environment of the Czech and Slovak Republic to enter the localities of populations of this species, and for manipulation with the species.

Study populations Demographic data were collected from nine populations of L. sibirica in the Czech Republic and two populations in the Slovak Republic (Table 1). The study populations in the Czech Republic are located in three regions. The first one (containing five local populations) is in northern Bohemia near the village Jestřebí (50°36′8″–50°36′22″ N, 14°36′51″–14°37′21″ E). The second one (containing three local populations) is in central Bohemia near the village Rečkov (50°29′38″– 50°30′3″ N, 14°51′56″–14°54′40″ E). The last population is in southern Bohemia in Pošumaví, by the fishpond ‘Olšina’ (48°47′34″ N, 14°6′7″ E). In the Slovak Republic, one population was located in NP Slovenský raj. It is an extensive population along the river Hnilec (48°52′32″ N, 20°15′46″ E). The second population is in Salvatorské lúky, approximately 25 km west of the town Prešov (49°2′45″ N, 20°56′30″ E). The population size was estimated in 2006. In the case of small populations (<500 adult individuals), the population size was the sum of all adult individuals in the locality. In the case of large populations (>500 adult individuals), the population size was estimated as an area of the population multiplied by the average number of adult individuals in ten 1-m2 plots. The plots were located throughout the locality to cover the different densities of adult individuals in the proportions in which they occur at the localities (Table 1).

24 Chapter I

______to states (CZ and SK) and regions and regions (CZ and SK) to states (none,September) + Shrubandremoval tree . The populations were divided according . The populations were divided according Ligularia sibirica Ligularia udied populations and localities of and populations udied

e CZ N/4 323 e CZ nitrogen-rich None č (S) (S) (N) (SR) SK SR 50,000 (SR) SK ~ nitrogen-poor None e CZ N/3 193 e CZ None nitrogen-rich (C) (C) č r CZ N/5 123 nitrogen-rich Mosaic mowing ů ka CZ C/3 ~ 7,000 nitrogen-poor MowingSeptember ní dv ní č č (P) (P) kov CZ C/1 ~ 10,000 nitrogen-poor Mosaic mowing(June, September) č Rákosina u Robe Rákosina Slune Vojenské louky Baronský rybník u Robe Louky CZ N/1 CZ N/2 126 243 nitrogen-poor nitrogen-poor Mosaic mowing(none, June,September) Shrub + andremoval tree Mowing(June) + Shrub and tree removal Olšina Olšina (SK) Republic Slovak raj NP Slovenský S/1 CZ 224 nitrogen-poor removal tree and Shrub Prešov Table 1: Characteristics of the st Characteristics Table 1: northern Bohemia Bohemia northern (N, C, S, SRand P) ______Population ______(CZ) Czech Republic Abbrev. size . pop. Estimated type Habitat Management of adults2006 in Habitat type Management Re LoukyNV u Kloko CZ C/2 ~ 10,000 nitrogen-rich None southern Bohemia Bohemia southern Salvatorské lúky ______P/1 SK ~ 2,000 nitrogen-poor None central Bohemia central Bohemia

25 Chapter I

To describe habitat type in the studied localities of L. sibirica we recorded one to three phytosociological relevés in 5 × 5 m plots in each locality in such a way that the heterogeneity of the locality was covered. Each vascular plant species and bryophyte in each relevé was assigned an Ellenberg indicator value for light, moisture, soil reaction and nitrogen availability (Ellenberg et al. 1992) if the value was available. The differences between relevés within a single locality were quite low, thus we used only one representative phytosociological relevé at each locality and for those we thereafter calculated the mean of each Ellenberg indicator value. Based on the Ellenberg indicator values for nitrogen availability, we divided the study localities into two groups: nitrogen- poor and nitrogen-rich (Table 1; Fig. S1 in Electronic Supplementary Material (hereafter ESM)). We also checked other Ellenberg indicator values (i.e. moisture, soil reaction and light) for their correlation with nitrogen availability. There was a strong negative correlation of Ellenberg indicator values for nitrogen availability and light (r=-0.79; P<0.01) (Table S1 in ESM). We used the Ellenberg indicator values for nitrogen availability to describe habitat conditions because they are widely and successfully used in this respect (Diekmann 2003; Käfer & Witte 2004). They were mainly developed on the basis of field experience and they reflect the ecological behaviour of species. Therefore, the average indicator values do not refer to habitat conditions at a specific moment, but present integration of habitat conditions over time (Ellenberg et al. 1992; Schaffers & Sýkora 2000). Thanks to this we suggest that they are more informative than direct measurements of nutrient availability in the soil because direct measurements reflect current status and may strongly vary over time (correlation between Ellenberg N and N content in the soil in our system is r=0.21, unpubl. data). Currently, localities of L. sibirica are under variable management regimes. Entire or in some cases only part of some localities are mown at different times (early June, late September), while other localities remain unmanaged (Table 1).

Demographic Data Collection and Construction of Transition Matrices Detailed demographic observations were conducted over four years from 2005 to 2008 and the life cycle of L. sibirica was divided into seven stages: (1) seed bank first year, (2) seed bank second year, (3) seedlings, (4) juveniles, (5) vegetative adults (6) fertile adults and (7) dormant adults (Fig. 1). A ‘seed bank’ is comprised of seeds that are in an enforced dormant state on or directly below the soil surface. To estimate the survival of seeds in the seed banks, we buried 10 nylon bags, each containing 50 seeds, in each study locality in autumn 2006 to a 5-cm depth below soil surface to prevent the seeds from germinating. In the autumn of 2007 and 2008, we excavated five seed bags per year and locality and tested the seeds for viability. The viability of the seeds was checked by cutting them to see if they contained firm endosperms (see Baskin & Baskin 1998). Seeds without firm endosperms were classified as dead.

26 Chapter I

a57 a16 a36 a56 a67 1 2 3 4 5 6 7

a21 a32 a43 a54 a65 a76

a44 a55 a66 a77

a31 a53 a75

Fig.1: Life-cycle diagram of the seven stages of Ligularia sibirica: (1) seed bank first year, (2) seed bank second year, (3) seedlings, (4) juveniles, (5) vegetative adults, (6) fertile adults and (7) dormant adult individuals

To assess seed germination and seedling establishment, we established five sowing plots and five control plots (50 × 50 cm) at each locality, scattered in such a way as to cover the heterogeneity of the locality. In the time of seed production (autumn) we added 50 developed seeds into each sowing plot. Fifty seeds were used to have enough seedlings in the sowing plot while avoiding competition between established seedlings. The control plots were established to estimate background germination from natural seed rain. The number of seedlings germinating in the control plots was then subtracted from the number of seedlings germinating in the sowing plots to estimate germination of the sown seeds (on average 4.16±7.22 seedlings, mean ± s.d., were detected in the control plots). Seedling emergence in the seedling recruitment plots and in the controls was monitored in autumn in subsequent years to check for possible delayed germination. The sowing plots represented different microhabitat types within the locality (e.g. bare ground, ground with litter, ground with bryophytes, ground with vascular plants and their combination). Then, we estimated the percentage cover of each microhabitat type in the whole locality. The resulting seed germination in a locality was estimated as seed germination in the sowing plots of different types weighted by the proportion of the given microhabitat in the whole locality. In the case of no germination in the sowing plots (i.e. localities CZ N/3, CZ N/4 and CZ N/5), we also estimated the germination rate by counting the number of new seedlings in the locality where we had marked all seedlings and juveniles the year before. We then calculated the ratio between the number of new seedlings and the number of fertile adult individuals the year before and directly obtained information on the number of seedlings per fertile adult individual. A similar approach was described by Lehtilä et al. (2006). The stage of ‘seedlings’ was defined as newly established individuals that developed directly after germination of the seeds. To estimate the survival rate of the seedlings, we followed the

27 Chapter I seedlings in the sowing plots and seedlings that newly occurred in the locality (as described above). We marked and followed at least 50 seedlings in each locality and year, when available. The stage ‘juvenile’ was defined as a seedling older than one year, with a maximum of two leaves and with the width of the largest leaf <2.5 cm. Individuals may stay in this stage for several years. All larger individuals were placed in the vegetative adult stage. To estimate the survival rate of juveniles, we marked and followed at least 50 individuals in this stage in each locality and year, when available. The classification of ‘adult’ individuals was based on the reproductive state (vegetative or fertile). Further division of the vegetative and fertile plants into different size categories was not possible as the main differences in size of the plants are at the between population level. Size differences thus reflect differences in habitat type (nutrient status of the soil) rather than life stage of the plant. Based on common garden observations a single adult individual (genet) of L. sibirica may consist of 1 to 15 rosettes in a dense cluster that is connected by up to 10-cm-long underground rhizomes. Because L. sibirica is a critically endangered species, we studied the length of the underground rhizomes by excavating adult individuals grown in the common garden experiment and not directly in the field. The same length of rhizomes was observed across all populations (pers. observation). We used this knowledge to distinguish adult individuals in the field and we assumed that the studied individuals corresponded to genets and thus the whole life cycle of the species was based on the level of genet. In small populations (<500 adult individuals), we marked and followed all adult individuals. In large populations (>500 adult individuals), we established two to four permanent plots in such a way that the heterogeneity of the locality was covered and altogether we marked at least 50 adult individuals in each life stage category (vegetative adults and fertile adults) to obtain reliable estimates of transition probabilities between stages (Münzbergová & Ehrlén 2005). If there were fewer than 50 living marked individuals per stage at the time of the census we additionally marked new individuals to maintain the minimum number of individuals needed per stage for the next transition interval. Furthermore, because we recorded very low mortality in the first transition interval, we additionally marked 50 adult individuals per population, when available, to increase the accuracy of estimating the probability of adult mortality. The new individuals were chosen near the permanent plots so that they could be easily relocated. Observations of the marked individuals were made twice per year. During the flowering time (mid-July to the end of August), we recorded the survival of each marked individual. For adult individuals, we recorded the number of rosettes, the number of leaves and maximum leaf width. For fertile adult individuals, we additionally recorded the number of flowering stems and the number of flower heads per inflorescence. During seed maturation (September), we assessed the number of developed undamaged seeds per flower head on 20 randomly selected fertile adult

28 Chapter I individuals at the upper, middle and lower parts of the inflorescence. We estimated the number of seeds in different parts of the inflorescence because the inflorescence is very long (23.15±11.17 cm; mean ± s.d.), and we assumed that the position within the inflorescence may effect the number of seeds per flower head. These data were used to estimate the seed production of all fertile adult individuals in each locality. Separately, we also noted the number of empty seeds and the number of seeds damaged by seed herbivores. We divided the production of seeds into two groups to describe generative reproduction. The first group was the seeds that germinated in the first year, which represented the direct transition from fertile adult individuals to seedlings. The information on germination thus means number of seeds that survive from autumn to spring when they germinate, establish and survive until the census, which was done annually at the time of flowering. The remaining seeds were expected to go to the seed bank and represented the transition from fertile adult individuals to the seed bank. The transition from fertile plant to seed bank thus includes all non-germinated seeds. That not all the seeds survive in the seed bank over the field season is taken into account in the transition from the seed bank to seedlings (Table 2). In this way, the approach used to estimate seed bank survival is correct when considered within the whole life cycle. However, the contribution of fertile plants to 1st year seed bank is overestimated and survival of seeds in the seed bank is underestimated (see Emery & Gross 2005). These transitions should thus always be interpreted together. This transition consists of survival of seeds in the seed bank and probability of germination. To get to the seed bank of the 2nd year, the seeds need to survive and not germinate the year before. Transition to the seed bank of the 2nd year was thus calculated as seed bank survival multiplied by the probability of non-germination. Probability of non-germination was needed in the calculation because the seed bank survival is estimated in bags, when the seeds are buried too deep to germinate. In contrast, in the field they are closer to the surface and can germinate. The experiment with seed burial in the nylon bags showed very low survival of seeds in the seed bank until the second year and thus we assume that no seeds survival until the third year. Therefore, we only considered seed banks in the first and second years in our models (Table 2; Table S2 in ESM). The stage ‘dormant adults’ was defined as individuals in a dormant state, without any organs visible above ground that reappeared in the following year or over two years. The probability to be dormant over a single year was 5.08% on average, and the probability of being dormant over two years was 1.11%. No plant was observed to be dormant over more than two years (result based on our preliminary observation of plants that started already in 2002). Therefore, an individual that was dormant for more than two years was considered dead. The transition to (a75, a76) dormancy was known only for the first two transition intervals and transition out (a57, a67) of dormancy were known only for two latter transition intervals. We used the mean of these two values for each of the three transition periods to deal with the missing values. Specifically, transition to dormancy was calculated as a probability of dormancy * probability of disappearance (dormancy + mortality) in

29 Chapter I

) fd G ______– probability – of v + mean – growth of seedlings of seedlings – growth vd js G

vd fd G G – survival of vegetative adults, adults, of vegetative – survival – growth of dormant adults to to adults of dormant – growth vv fd mean mean 1 - (mean

) Ger

– probability of dormancy of fertile adults, – probabilityof dormancy fertile adults, of fd f Ger Ger – survival of fertile adults, D adults, fertile of – survival

* D ff

f vf * * (1 * (1 - ff S S – survival in the seed bank, G sb D vd S * D R fv v vv D G

– growth of juveniles to vegetative adults, S adults, to vegetative of juveniles – growth of seeds in the field, S vj S

vj – growth of dormant adults to vegetative adults, G adults, to vegetative adults of dormant – growth jj tive adults (mortalityD + dormancy),tive S G vd

js vs G G

– survival of juveniles, G juveniles, of survival – jj individual, Ger – germination individual, Ger – germination – retrogression from fertile adults to vegetative adults, S adults, vegetative to adults fertile from – retrogression vf year seedlings juveniles vegetative adults fertile adults dormant adults

nd Ger Ger *

sb ile adults (mortalitydormancy), + G

Ligularia sibirica Ligularia – probability of disappearance of vegeta

vd ) year 2 seed bank st Ger Ger S Ger *

* (1 - * (1 sb sb S S – growth of seedlings to vegetative adults, S adults, vegetative to of seedlings – growth vs year year st nd – growth of vegetative adults to fertile adults, R adults, fertile to adults vegetative of growth – – probabilityofof disappearancefert fv fd dormancy of vegetative adults, D adults, of vegetative dormancy dormant adults ______seed bank 1 seed bank fertile adults fertile adults ______by a fertile adult number of seeds produced S – average vegetative adults mean juveniles juveniles mean to juveniles, G to juveniles, seed bank 2 seed bank Table 2: Transition matrix model of of model matrix 2: Table Transition G D seedlings seedlings ______seed bank 1 bank seed

30 Chapter I the given size category. Probability of dormancy was a mean of proportion of dormant individuals from dead individuals in the given size category. Probability of disappearance in the given size category was then the sum of individuals that disappeared in the given size category in the given time period (dead + dormant) divided by total sum of individuals in the given size category (Table 2).

Data Analyses To investigate the differences in vital rates within the life cycle of L. sibirica between populations and years, we performed a set of regression analyses. First, we tested the effect of year, population and position of flower head within the inflorescence (upper, middle, lower parts) and their interactions on the number of initiated and developed seeds (including those subsequently damaged by herbivores) using general linear models with normal distribution. Then, we tested the effect of these independent variables on the proportion of damaged seeds using general linear models with Poisson distribution (number of damaged seeds used as dependent variable and total number of seeds used as a covariate). We also tested the effects of population and year on specific vital rates that were included directly in the transition matrices: probability of stasis (i.e. individual stays in the same stage), growth (i.e. transition to the next stage), and flowering and mortality using logistic regression. These parameters were determined in individuals of different stages. To remove this effect, we used the stage in the previous year as the independent variable in the tests. All tests were carried out using the statistical program S-Plus (2000). Demographic data were examined using transition matrix models (Caswell 2001). To describe the performance of each population, we calculated the population growth rate (Caswell 2001) and elasticity (contribution of each matrix element to population growth rate, de Kroon et al. 2000) for each population and transition interval separately. We calculated 95% confidence intervals to estimate the reliability of the estimates of population growth rates and elasticity (Alvarez-Buylla & Slatkin 1994) by bootstrapping the original data used to derive the transition matrices 10,000 times, as suggested by Efron and Tibshirani (1994) and Caswell (2001) using a MATLAB script developed by Münzbergová (2006). To summarize the information on population growth rate over populations and years, we used the stochastic simulation approach suggested by Caswell (2001) and Sletvold and Rydgren (2007). We drew a sequence of matrices for each set of matrices. Each matrix from the set was drawn at random and with equal probability and we simulated population growth using this matrix sequence. Each simulation was performed for 10,000 one-year intervals. The simulations were performed using a MATLAB script developed in a previous study (Münzbergová 2005). We ran the same stochastic simulations for the above-described bootstrapped matrices and thus were able to construct a 95% confidence interval for the stochastic population growth rate.

31 Chapter I

To summarize the information on the elasticity of single matrix elements, we divided the single elements into growth (G), stasis (L) and fecundity (F), following Silvertown and Franco (1993), and summed the elasticity values within these three categories. We, however, also compared the contributions of single matrix elements to get more detailed insights into the differences. A life-table response experiment (LTRE) with a fixed factorial design was conducted to examine the effects of population and year on population growth rate, as suggested by Caswell (2001). To test the differences between small and large populations and populations from nutrient- rich and nutrient-poor habitats, we selected only a subset of populations to have the same number of populations (four) in each category. To test the differences in population size we included all populations in northern Bohemia except CZ N/1, all populations in central Bohemia and SK P/1. To test the differences in habitats we included all northern and central Bohemian populations. We calculated the mean matrix (i.e. across all populations and years) and its growth rate. Differences in the growth rates of the matrices of each type from the growth rate of the mean matrix were calculated (see Münzbergová 2007). The LTRE analysis indicates the contribution of each life cycle transition to the differences between different levels of each factor (population, year). Important life-cycle transitions are those with large positive contributions at some factor levels and large negative contributions at others. Similarly to ANOVA, the mean of the treatment effects is zero (Caswell 2001). The significance of the LTRE was estimated using the permutation test as described in Münzbergová (2007), using 10,000 permutations. To calculate the life span, we used the method suggested by Cochran and Ellner (1992), which derives age-based measures from stage-structured models. The resulting conditional life span is defined as the average age of death for individuals that survive to a certain stage (Cochran & Ellner 1992). To represent the population, we decided to use the life span of the mature stage that has the highest longevity values, i.e. fertile individuals (see Ehrlén & Lehtilä 2002). The method of Cochran and Ellner (1992) only works with a single matrix. To obtain conditional total life span data over all studied years in each population, we computed the harmonic mean of the life spans for all matrices representing different years within one population as a measure of the average life span for the species in the locality. We then calculated the arithmetic mean of all population estimates to obtain one value for the studied species (see Ehrlén & Lehtilä 2002). To calculate the probability of extinction of each population, we used the number of individuals in each population from the first year of the demographic observations. Because the number of seeds in the seed bank and the number of adult individuals in dormancy was not known, we complemented this missing information according to the mean stable stage distribution based on the transition matrices. Projections of the obtained population vector were made over 25, 50, 100 and 200 years. This approach was repeated 500 times using a MATLAB script. The number of individuals in each stage in each step (year) was replaced by a number drawn from a Poisson

32 Chapter I distribution with a mean corresponding to the number of individuals in that stage. In this way, we simulated demographic stochasticity. All analyses were performed using MATLAB version 7.3.0.267 (Mathworks Inc. 2006).

Results

Specific Vital Rates There were strong differences between populations, years and position of flower head in the inflorescence (upper, middle, lower parts), in the number of initiated and developed seeds and in the proportion of damaged seeds with population having the strongest effect (Table 3). The highest number of seeds was initiated and developed in populations growing in nitrogen-rich habitats (i.e. CZ N/3, CZ N/4 and CZ C/2) and in the lower part of the inflorescences (see Table S3 in ESM). The highest number of herbivore-damaged seeds was found in the middle part of the inflorescence. There were significant differences in the probabilities of stasis, growth, flowering and mortality between single populations, years and stages in the previous year. The only exception was the probability of flowering, which did not differ between years. There were also many significant interactions between populations, years and stages in the previous year. Most of the variability was explained by population (Table 4). When comparing the populations, the most variation was found for the probability of flowering. The highest probability of flowering was found in populations CZ N/3 and CZ C/2, which grow in nitrogen-rich habitats, and the lowest probabilities of flowering were found in the small population of CZ N/5, which grows in a nitrogen-rich habitat, and in the large population of CZ C/1, which grows in a nitrogen-poor habitat.

Table 3: Regression analysis of population, year, position in inflorescence and their interactions in relation to the number of initiated and developed seeds and on the proportion of damaged seeds. The number of initiated and developed seeds were tested using general linear models with normal distribution and the proportion of damaged seeds was tested using general linear models with Poisson distribution. D.f. error = 1979. Significance levels: * – P<0.05, ** – P<0.01, *** – P<0.001

Nr. of initiated seeds Nr. of developed Proportion of

seeds damaged seeds % explained % explained % explained d.f. P P P variability variability variability Population 10 *** 27.80 *** 23.47 *** 19.01 Year 2 n.s. - *** 1.94 *** 1.25 Position in inflorescence 2 *** 9.27 *** 7.27 ** 0.79 Population × year 20 *** 2.76 *** 2.97 *** 6.21 Population × position in 20 *** 2.76 *** 3.78 *** 4.01 inflorescence Year × position in 4 * 0.29 *** 0.71 *** 0.96 inflorescence Population × year × position 40 *** 4.41 *** 4.82 *** 7.42 in inflorescence

33 Chapter I

______% explained % explained variability 1.50 1.50 1.32 1.32 5.42 5.42 2.37 2.37 4.30 4.30 0.24 0.24 1.56 1.56 P

*** *** *** *** *** ** ** *** Probabilitymortality of

lity of growth, d.f. error = 7546 for the d.f. for the = error of growth, 7546 lity – P<0.001 the same stage, the probabilityof growth, % explained % explained d.f. variability 9.20 10 10 9.20 2.28 1 1 2.28 0.72 20 20 0.72 0.48 10 10 0.48 0.69 2 2 0.69 P

*** *** *** *** *** * – P<0.05, ** P<0.01, *** the probabilityof survival in d.f. error = 5893 for the probabi d.f. error = 5893 for Probability of flowering of Probability % explained % explained d.f. variability 1.95 10 10 1.95 0.28 2 n.s. - - 2 n.s. 2 0.28 2.45 1 1 2.45 0.48 20 20 0.48 2.85 10 10 2.85 P

*** *** *** * *** rtality. Significance levels: rtality. Significance survival in the same stage, in the same survival eir interactions in relationeir interactions to % explained % explained d.f. variability 1.49 10 10 1.49 0.07 2 2 0.07 0.79 1 1 0.79 0.42 20 20 0.42 5.26 10 10 5.26 0.15 2 n.s. - - 2 n.s. 2 0.15 0.54 18 n.s. - 20 n.s. - 20 20 - n.s. 20 - n.s. 18 0.54 r the probability of mo of r the probability P

*** * *** ** ** *** *** *** d.f. error = 7300 for the probability of d.f. error = 7300 for the probability ______d.f. Survivalsame stage in the Probabilitygrowth of probability of flowering, d.f. error = 5893 fo 5893 flowering, d.f. error = of probability flowering and mortality. Table 4: regression Logistic analysis and of population, th year ______10 Population Year 2 Stage in previous year Stage in previous 1 Population × year year × Population 20 Population × stage stage × Population 10 in previousyear year in previous Year × stage 2 Population × year × × year × Population 20 in previousyear ______

34 Chapter I

Population Performance There was a high variation in population growth rate between the populations and a rather small variation between the different years. Out of 33 transition matrices (see Table S4 in ESM), eight had a population growth rate significantly lower than 1 and eleven had a population growth rate significantly higher than 1. None of the others differed significantly from 1 (see Table S5 in ESM). The stochastic population growth rates that combined all transition matrices within each population differed between the populations. Out of 11 populations, one had a stochastic population growth rate significantly lower than 1 and eight had a stochastic population growth rate significantly higher than 1 (Fig. 2). This result did not differ when we estimated the stochastic population growth rate with zero germination probabilities for populations CZ N/3-N/5 (0 germination was found in the germination plots and the germination was thus estimated as the number of natural seedlings per number of flowering plants, see Methods, results not shown). The analysis of elasticity indicated that the most important transition in all populations was survival of the vegetative adults (transition a55). This transition was followed by all growth transitions (transitions a54, a65), survival of fertile individuals (transition a66) and generative reproduction (transition a36), (see Fig. S2 in ESM). This means that survival of the adult individuals was much more important for the population growth rate than generative reproduction. When looking at the position of populations in the G-L-F triangle based on elasticity of growth, stasis and reproduction (Silvertown & Franco 1993; Silvertown et al. 1996), several trends emerged. The population growth rate increased as stasis (L) decreased and fecundity (F) and growth (G) increased. Small populations with a low growth rate growing in nitrogen-rich habitats were placed close to the L = 1 vertex of the triangle. The value of stasis was far larger than the values for growth and fecundity. Populations growing in nitrogen-poor habitats were placed close together in the middle of the triangle’s axis L. Contrary to populations with low growth rates, two small populations growing in nitrogen-poor habitats with larger fecundity and growth values than the value for stasis had a high growth rate (Fig. 3). The LTRE analysis indicated a significant negative contribution of the transition interval 2006–2007 and a strong but non-significant positive contribution of the transition interval 2005– 2006 to the overall population growth rate (Table 5a). There was a highly significant positive contribution of the currently increasing CZ N/1 population and a significant negative contribution of populations CZ C/1, CZ S/1 and SK SR (Table 5b) to overall population growth rates. All populations with significant contributions were growing in nitrogen-poor habitats. When decomposed into single matrix elements, the variation in population growth rate between years was mainly significantly affected by survival of the vegetative adults (transition a55) and fertile adults

(transition a66), generative reproduction (transition a36) and retrogression of fertile adults to vegetative adults (transition a56). The variation in population growth rate between the different

35 Chapter I

1.8

1.6

1.4 λ 1.2

1

0.8

/1 /2 /3 /4 /5 /1 /2 /3 /1 R /1 N N N N N C C C S S P Z Z Z Z Z Z Z Z Z K K C C C C C C C C C S S

Fig. 2: Stochastic population growth rate of Ligularia sibirica populations with 95% confidence intervals. The horizontal line indicate population growth rate (λ) = 1, i.e. stable population. See Table 1. for abbreviations of the study localities

Fig. 3: Triangular G-L-F ordination of the 11 elasticity matrices of Ligularia sibirica. G – growth, L – stasis, F – fecundity. See Table 1 for abbreviations of the study localities

36 Chapter I

Table 5: Contribution of year (a) and population (b) to the observed variation in population growth rates of Ligularia sibirica using life-table response experiments (LTREs). The significance values were determined using a permutation test and indicate that the contributions are significantly different from those expected at random. Significance levels: * – P<0.05, ** – P<0.01, *** – P<0.001

(a) Contribution P Year 2005–2006 0.12 n.s. 2006–2007 -0.10 * 2007–2008 -0.02 n.s.

(b) Contribution P Population CZ N/1 0.78 *** CZ N/2 0.24 n.s. CZ N/3 0.05 n.s. CZ N/4 0.04 n.s. CZ N/5 0.02 n.s. CZ C/1 -0.37 *** CZ C/2 -0.25 n.s. CZ C/3 -0.11 n.s. CZ S/1 -0.12 ** SK SR -0.17 *** SK P/1 -0.11 n.s.

populations was significantly affected by all transitions, except for the rare transitions a43, a53 and a67. The highest positive contribution between population variations was that of generative reproduction (transition a36) in population CZ N/1. Other high positive contributions were growth transitions (transition a54) in population CZ N/5, (transition a65) in populations CZ N/2, CZ N/3, CZ

C/2, CZ C/3 and survival of the vegetative adults (transition a55) in populations CZ C/1, SK SR.

High negative contributions were growth transitions (transition a54) in population CZ N/4,

(transition a65) in populations CZ N/5, CZ C/1, SK P/1 and survival of the vegetative adults

(transition a55) in populations CZ C/2, survival of the fertile adults (transition a66) in populations CZ S/1 (see Table S6 in ESM). Results of LTRE analyses comparing the populations divided into two groups depending on their population size and habitat type showed no significant differences between any of the groups (P>0.05 in both cases). When decomposed into single matrix elements, the variation in population growth rate between population size was mainly significantly affected by growth of vegetative adults to fertile adults (transition a65) and survival of the vegetative adults (transition a55) and fertile adults (transition a66). Small populations had positive contributions of transitions a65, a66, negative contribution of transition a55 and vice versa in large populations. The variation in population

37 Chapter I growth rate between the different habitat types was mainly significantly affected by survival of juveniles (transition a44), vegetative adults (transition a55), fertile adults (transition a66) and by transition from vegetative adults to dormant adults (transition a75). Nitrogen-rich populations had positive contribution of transitions a66, a75, negative contribution of transitions a44, a55 and vice- versa in nitrogen-poor populations (Fig. 4).

Life Span The mean life span of L. sibirica was 61.3 years. When looking at the habitat difference, populations growing in nitrogen-poor habitats had a mean life span of 77.6 years (range: 20.4– 135.2) and populations growing in nitrogen-rich habitats had a mean life span of 30.9 years (range: 13.4–49.9).

Risk of Extinction Analysis of extinction probabilities showed differences between populations growing in different habitats. Two small populations growing in nitrogen-rich habitats (CZ N/3, CZ N/4) will become extinct in 50 years with a 20% probability and in 100 years with a 95% probability. Furthermore, another small population (CZ N/5) also growing in a nitrogen-rich habitat will become extinct in 25 years with an 80% probability and in 50 years with a 100% probability. A zero extinction probability within 200 years was found for both large and small populations of L. sibirica growing in nitrogen-poor habitats and for one large population growing in a nitrogen-rich habitat (CZ C/2).

Discussion Results showed that most studied populations of Ligularia sibirica in the Czech and Slovak Republic are performing well in terms of population growth rate and that only the small populations growing in nitrogen-rich habitats showed a decreasing trend. Analyses also showed significant temporal and spatial variation. Whereas a temporal difference in population growth rate was mainly driven by survival of adult individuals, spatial variation was largely driven mainly by high generative reproduction of one small currently expanding population. The population growth rate of small populations was driven mainly by the growth of adult individuals and flowering. The population growth rate of populations growing in nitrogen-poor habitats was mainly driven by survival of vegetative individuals. Our results also indicate that all populations can be characterized by low adult mortality and long-lived individuals, and some populations also showed features of remnant populations. Detailed demographic data are thus needed to understand the long-term prospects of these populations. They may detect negative changes in the populations long before the changes are visible in the number of flowering plants and may thus support the initiation of management actions towards population restoration before it is too late. Analyses of vital rates showed highly significant differences between populations and between years in survival, growth,

38 Chapter I

a)

b)

Fig. 4: The positive and negative contributions of matrix elements to the observed variation in population growth rate between different a population size and b habitat type of Ligularia sibirica using life table response experiment. The significance values were determined using a permutation test and indicate that the contributions are significantly different from those

39 Chapter I flowering and mortality of individuals and in number of initiated, developed seeds and proportion of damaged seeds. The intensity of pre-dispersal seed herbivory was the lowest in 2006. When comparing populations, the highest intensity of pre-dispersal seed herbivory was found in a large population growing in nitrogen-poor habitat SK SR. This population also had the lowest seed production (number of developed undamaged seeds). From our results we cannot say which of these two factors (i.e. population size, habitat type) influenced the intensity of herbivory most, but the effect of both were already previously reported (Östergård & Ehrlén 2005; Ågren et al. 2008). Similar strong differences between years and populations were also apparent from the LTRE analyses, indicating high temporal and spatial variations in population growth rates. The temporal variation is most likely caused by a change in climatic conditions (Fréville et al. 2004) or, but less likely, by successional changes in vegetation structure (Oostermeijer et al. 1996) and the intensity of pre-dispersal seed predation (Ehrlén 1996). In our case, temporal variation was mainly driven by the transition interval 2006–2007. In 2006, there were a very low number of developed seeds. This might have been due to the extremely dry July, as well as the extremely wet September in 2006, when compared with the long-term average (Český hydrometeorologický ústav 2010; Slovenský hydrometeorologický ústav 2010). The spatial variation could be caused by the differences in population growth rate between population size (small and large populations) and habitat type (nitrogen-poor and nitrogen-rich habitats). According to the LTRE analysis, the most deviating population was a small population, which is currently increasing and grows in a nitrogen-poor habitat (CZ N/1). There was high seedling recruitment in this population because of the presence of suitable micro-environmental conditions for germination in terms of a well-developed moss layer. The moss layer was identified as the most suitable microhabitat for germination based on a detailed comparison of the results of the sowing experiment described above (not shown). Positive effects of the moss layer were also demonstrated in previous studies (Overbeck et al. 2003), while other studies indicated that the moss layer can also inhibit seedling establishment and growth (Kottorová & Lepš 1999; Rydin & Jeglum 2006). In contrast to population CZ N/1 (small and currently expanding population growing in nutrient-poor habitat), the other populations growing in nitrogen-poor habitats (i.e. CZ S/1, CZ C/1) contributed negatively to the overall population growth rates. This was due to the negative contributions of the transition probabilities of the growth of vegetative adult individuals to fertile adult individuals and the survival of fertile adults. The low flowering rate observed was probably caused by missing or unsuitable habitat management actions at these localities in terms of long- term early mowing (Šmídová 2009, unpubl. data). Our results suggest that seed survival in the seed bank only contributes to a little of the variation in population growth rate. This is because the seed bank of L. sibirica is small when compared with seed production (transient seed bank type II, according to Thompson & Grime

40 Chapter I

1979). In the nitrogen-poor habitats, most seeds germinate or decay in the first year after production, resulting in almost no seeds in the seed bank. In contrast, the seed bank in nitrogen-rich habitats can survive for two subsequent years. This is due to the low groundwater table in these localities. The contribution of the seed bank in these habitats to population growth rates is, however, very low because of the very low seedling survival rate in these localities. The low rates of seedling survival in the nitrogen-rich habitats were probably due to competition for light with dense tussocks and because of the low groundwater table (see Schopp-Guth et al. 1994 for a similar pattern). In contrast to the retrospective (LTRE) analysis, the prospective (elasticity) analysis indicated differences between habitat types, while the differences between population sizes were indistinct. For populations growing in nitrogen-poor habitats, growth of seedlings, survival and the growth of juveniles and vegetative adults, and also fecundity, had the highest elasticity. Populations growing in nitrogen-rich habitats had the highest elasticity for the survival of vegetative and flowering adult individuals, indicating that small changes in these transitions would have strong effects on the population growth rate. A high importance of the survival of adult individuals was found for many perennial wetland plant species, for instance Gentiana pneumonanthe and Scorzonera humilis (Oostermeijer et al. 1996; Colling & Matthies 2006). This finding could be explained by the ability of L. sibirica to store resources in a short rhizome that help the plant to buffer temporal variations in its environment. The importance of adult dormancy slightly increased in populations growing in nitrogen-rich habitats. Adult dormancy is mainly known for the Orchidaceae family (reviewed in Shefferson 2009), and it is only rarely considered in studies on complete population dynamics (e.g. Ehrlén 1995; Oostermeijer et al. 1996). As in other species, adult dormancy in L. sibirica could be a response to unfavourable environmental conditions (the so-called ‘storage effect’; Higgins et al. 2000). When we compared both prospective and retrospective analyses, we obtained slightly inconsistent insights. The elasticity analyses identified stasis as the most important transition for population growth rate in populations growing in nitrogen-rich habitats. In contrast, the LTRE analyses showed that stasis negatively contributed to the population growth rate of L. sibirica. This finding is consistent with the conclusion of Lehtila et al. (2006) for decreasing late-successional populations of Primula veris. It suggests that the best management practice for these populations is to increase the stasis of individuals and to support the growth of individuals even though the elasticities were low. For the populations growing in nitrogen-poor habitats the results of prospective and retrospective analyses were in accordance. They indicated a positive contribution of stasis to the population growth rate and a high importance of fecundity. The extinction probability for both the small and large populations growing in nitrogen-poor habitats, as well as for the large population growing in the nitrogen-rich habitat (CZ C/2), was zero, even when calculated for the next 200 years. This result may be due to the long life span and slow

41 Chapter I growth and stability of these populations, which seem to be typical for many species occurring in habitats with adverse growth conditions (Eriksson 1996; García 2003). For the small populations growing in nitrogen-rich habitats, CZ N/3–4 and CZ N/5, the predicted probability of extinction in 100 years was 95% and within 25 years it was 80%. These three small populations have been growing for more than 50 years in deteriorating environmental conditions due to melioration (Hendrych 2003) and the expected life span of the individuals is thus very low. The above-mentioned L. sibirica populations clearly showed characteristics of remnant populations (CZ N/3–5). These populations can persist for an extended period of time under severe environmental conditions (Eriksson 1996, 2000) with almost no reproduction thanks to the longevity of adult individuals, their low mortality rates (Oostermeijer et al. 1994; Colling & Matthies 2006) and also, in the case of L. sibirica, adult dormancy. This leads to a strong delay between the deterioration of habitat conditions and the reduction of population size (Colling & Matthies 2006). We are aware that the estimates of plant dormancy and seed bank longevity are based on short-term data, but, according to the results, these parts of the life cycle have very small effects on the population growth rate of L. sibirica. Furthermore, the analysis of extinction probabilities based on such short-term data compared with the longevity of the study species cannot give us accurate long-term predictions. Such a prediction, however, can identify potential extinction threats for populations in the near future (Ehrlén & Lehtilä 2002).

Conservation Implications The results of our study indicate that most of the studied populations of L. sibirica, both small and large, are not declining and that they have very low extinction probabilities. The one population that is currently declining and the other two populations that showed a tendency towards decline were small and occurred in nitrogen-rich habitats. The elasticity of stasis in these declining populations is very high. This, together with almost no reproduction, suggests that these populations are remnants. Increasing the population growth rate of these populations would thus require active management actions that should not only aim towards prolonging the life span of individuals, but should also support recruitment of new individuals into the population. High extinction rates of these populations are expected in 100 years. This demonstrates that the population viability analysis performed in this study provides useful information on population performance before the decline can be detected from population size. Our findings demonstrate that we should be careful with generalizing knowledge obtained from one population and applying it to others. When attempting this, we need to mainly focus on the identified critical parts of the life cycle as well as on habitat conditions because these factors were found to be the most important drivers of population performance, and, as such, they are therefore targets of specific management actions.

42 Chapter I

Acknowledgments We would like to thank the workers in the protected areas of the Czech and Slovak Republic in which L. sibirica occurs for providing us with information about the localities. We obtained permission from the Ministries of the Environment of the Czech and Slovak Republic to enter the localities of this species’ populations, and to manipulate with the species. This study was supported by grant MŠMT 2B06178 and also partly by MŠMT 0021620828 and AV0Z60050516.

References Ågren, J., Ehrlén, J. & Solbreck, C. (2008): Spatio-temporal variation in fruit production and seed predation in a perennial herb influenced by habitat quality and population size. J. Ecol., 96: 334–345. Alvarez-Buylla, E.R. & Slatkin, M. (1994): Finding confidence limits on population growth rates: three real examples revised. Ecol., 75: 255–260. Baskin, C.C. & Baskin, J.M. (1998): Seeds: ecology, biogeography, and evolution of dormancy and germination. Academic Press, San Diego Brussard, P.F. (1991): The role of ecology in biological conservation. Ecol. Appl., 1: 6–12. Caswell, H. (2001): Matrix population models, construction, analysis, and interpretation. Sinauer Associates, Sunderland Český hydrometeirologický ústav (2010): Informace o klimtu v České republice. Průměrná úhrn srážek a teplot v letech 2005-2008 ve srovnání s dlouhodobým normálem z let 1961-1990. Available at: http://old.chmi.cz/meteo/ok/infklim.html, accessed 25 November 2010 Cochran, M.E. & Ellner, S. (1992): Simple methods for calculating age-based life history parameters for stage-structured populations. Ecol. Monogr., 62: 345–364. Colling, G., Matthies, D. & Reckinger, C. (2002): Population structure and establishment of the threatened long-lived perennial Scorzonera humilis in relation to environment. J. Appl. Ecol., 39: 310–320. Colling, G. & Matthies, D. (2006): Effects of habitat deterioration on population dynamics and extinction risk of an endangered, long-lived perennial herb (Scorzonera humilis). J. Ecol., 94: 959–972. Council of European Communities (1992): Council Directive 92/43/EEC of 21 May on the conservation of natural habitats and of wild fauna and flora. Off. J. Eur. Communities 35: 7– 50. Crawley, M.J. (1989): Insect herbivores and plant population dynamics. Annu. Rev. Entomol., 34: 531–564. Dahlgren, J.P. & Ehrlén, J. (2009): Linking environmental variation to population dynamic of a forest herb. J. Ecol., 97: 666–674. De Kroon, H., van Groenendael, J. & Ehrlén, J. (2000): Elasticities: a review of methods and model limitations. Ecol., 81: 607–618. Diekmann, M. (2003): Species indicator values as an important tool in applied plant ecology – a review. Basic Appl. Ecol., 4: 493–506. Doust, L.L. (1981): Population dynamics and local specialization in a clonal perennial (Ranunculus repens): I. The dynamics of ramets in contrasting habitats. J. Ecol., 69: 743–755. Efron, B. & Tibshirani, R.J. (1994): Introduction to the bootstrap. Chapman and Hall, New York Ehrlén, J. (1995): Demography of the perennial herb Lathyrus vernus. II. Herbivory and population dynamics. J. Ecol., 83: 297–308. Ehrlén, J. (1996): Spatiotemporal variation in predispersal seed predation intensity. Oecologia, 108: 708–713. Ehrlén, J. & Lehtilä, K. (2002): How perennial are perennial plants? Oikos, 98: 308–322. Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W., & Paulißen, D. (1992): Zeigerwerte von Pflanzen in Mitteleuropa, Ed. 2. Scripta Geobot., 18 :1–258. Emery S.M. & Gross, K.L. (2005): Effects of timing of prescribed fire on the demography of an invasive plant, spotted knapweed Centaurea maculosa. J. Appl. Ecol., 42: 60–69.

43 Chapter I

Eriksson, O. (1996): Regional dynamics of plants: a review of evidence for remnant, source-sink and metapopulations. Oikos, 77: 248–258. Eriksson, O. (2000): Functional roles of remnant plant populations in communities and ecosystems. Global Ecol. Biogeogr., 9: 443–449. Feráková, V., Maglocký, Š. & Marhold, K. (2001): Červený zoznam papraďorastov a semenných rastlín Slovenska (december 2001). In Baláž, D., Marhold, K. & Urban, P. (eds) Červený zoznam rastlín a živočíchov Slovenska. Ochr. Prír., Banská Bystrica, 20: 48–81. Fréville, H., Colas, B., Riba, M., Caswell, H., Mignot, A., Imbert, E. & Olivieri, I. (2004): Spatial and temporal demographic variability in the endemic plant species Centaurea corymbosa (Asteraceae). Ecol., 85: 694–703. García, M.B. (2003): Demographic viability of a relict population of the critically endangered plant Borderea chouardii. Conser. Biol., 17: 1672–1680. Hegi, G. (1929): Illustrierte Flora von Mittel – Europa, Ed. 1. Vol. 6/2. München, pp. 796-799. Hegland, S.J., van Leeuwen, M. & Oostermeijer, J.G.B. (2001): Population structure of Salvia pratensis in relation to vegetation and management of Dutch dry floodplain grasslands. J. Appl. Ecol., 38: 1277–1289. Hendrych, R. (2003): Poznatky o druhu Ligularia sibirica v Čechách. Preslia, 75: 39–69. Higgins, S.I., Pickett, S.T.A. & Bond, W.J. (2000): Predicting extinction risk for plants: environmental stochasticity can save declining populations. Trends Ecol. Evol., 15: 516–520. Käfer, J. & Witte, M.J.P. (2004): Cover-weighted averaging of indicator values in vegetation analyses. J. Veg. Sci., 15: 647–652. Kottorová, I. & Lepš, J. (1999): Comparative ecology of seedling recruitment in an oligotrophic wet meadow. J. Veg. Sci., 10: 175–186. Kukk, Ü. (2003): The distribution of Ligularia sibirica (L.) Cass. in Estonia and changes in its population. Biul. Ogrodów Bot., 12: 11–22. Lehtilä, K., Syrjänen, K., Leimu, R., Garcia, M.B. & Ehrlén, J. (2006): Habitat change and demography of Primula veris: Identification of management targets. Conser. Biol., 20: 833– 843. Lennartsson, T. & Oostermeijer, J.G.B. (2001): Demographic variation and population viability in Gentianella campestris: effects of grassland management and environmental stochasticity. J. Ecol., 89: 451–463. Lindborg, R. & Ehrlén, J. (2002): Evaluating the extinction risk of a perennial herb: demographic data versus historical records. Conser. Biol., 16: 683–690. Matlab (2006): Matlab version 7.3.0.267. Mathworks Inc., Natick, MA Menges, E.S. (2000): Population viability analyses in plants: challenges and opportunities. Trends Ecol. Evol., 15: 51–56. Münzbergová, Z. (2005): Determinants of species rarity: population growth rates of species sharing the same habitat. Amer. J. Bot., 92: 1987–1994. Münzbergová, Z. (2006): Effect of population size on prospect of species survival. Folia Geobot., 41: 137–150. Münzbergová, Z. (2007): Population dynamics of diploid and hexaploid populations of a perennial herb. Ann. Bot., 100: 1259–1270. Münzbergová, Z. & Ehrlén, J. (2005): How best to collect demographic data for population viability analysis models. J. Appl. Ecol., 42: 1115–1120. Oostermeijer, J.G.B., Luijten, S.H. & den Nijs, J.C.M. (2003): Integrating demographic and genetic approaches in plant conservation. Biol. Conser., 113: 389–398. Oostermeijer, J.G.B., Van’t Veer, R. & den Nijs, J.C.M. (1994): Population structure of the rare, long-lived perennial Gentiana pneumonanthe in relation to vegetation and management in the Netherlands. J. Appl. Ecol., 31: 428–438. Oostermeijer, J.G.B., Brugman, M.L., de Boer, E.R. & den Nijs, H.C.M. (1996): Temporal and spatial variation in the demography of Gentiana pneumonanthe, a rare perennial herb. J. Ecol., 84: 153–166. Oostermeijer, J.G.B., Luijten, S.H., Křenová, Z.V. & den Nijs, J.C.M. (1998): Relationship between population and habitat characteristics and reproduction of the rare Gentiana pneumonanthe L. Conser. Biol., 12: 1042–1053.

44 Chapter I

Östergård, H. & Ehrlén, J. (2005): Among population variation in specialist and generalist seed predation – the importance of host plant distribution alternative hosts and environmental variation. Oikos, 111: 39–46. Overbeck, G., Kiehl, K. & Abs, C. (2003): Seedling recruitment of Succisella inflexa in fen meadows: Importance of seed and microsite availability. Appl. Veg. Sci., 6: 97–104. Primack, R.B. (2002): Essentials of conservation biology. Ed. 3. Sinauer Associates, Suderland Procházka, F. (ed) (2001): Černý a červený seznam cévnatých rostlin České republiky (stav v roce 2000). Příroda, Praha 18: 1–166. Reed, J.M., Mills, L.S., Dunning, J.B., Menges, E.S., McKelvey, K.S., Frye, R., Beissinger, S.R., Anstett, M.C.H. & Miller, P. (2002): Emerging issues in population viability analysis. Conser. Biol. 16: 7–19. Rydin, H. & Jeglum, J. (2006): The biology of peatlands. Oxford University Press, New York Schaffers, A.P. & Sýkora, K.V. (2000): Reliability of Ellenberg indicator values for moisture, nitrogen and soil reaction: a comparison with field measurements. J. Veg. Sci., 11: 225–244. Schemske, D.W., Husband, B.C., Ruckelhaus, M.H., Goodwillie, C., Parker, I.M. & Bishop, J.G. (1994): Evaluating approaches to conservation of rare and endangered plants. Ecol., 75: 584– 606. Schopp-Guth, A., Maas, D. & Pfadenhauer, J. (1994): Influence of management on the seed production and seed bank of calcareous fen species. J. Veg. Sci., 5: 569–578. Shefferson, R.P. (2009): The evolutionary ecology of vegetative dormancy in mature herbaceous perennial plants. J. Ecol., 97: 1000–1009. Silvertown, J. & Franco, M. (1993): Plant demography and habitat: a comparative approach. Plant Spec. Biol., 8: 67–73. Silvertown, J., Franco, M. & Menges, E. (1996): Interpretation of elasticity matrices as an aid to the management of plant populations for conservation. Conser. Biol. 10: 591–597. Slavík, B. (2004): Ligularia Cass. In Slavík, B. & Štěpánková, J. (eds) Květena České republiky, Vol. 7. Academia, Praha, pp 306–309. Sletvold, N. & Rydgren, K. (2007): Population dynamics in Digitalis purpurea: the interaction of disturbance and seed bank dynamics. J. Ecol., 95: 1346–1359. Slovenský hydrometeorologický ústav (2010): Informačný systém monitoringu životného prostredia. Čiastkový monitorovací. Available at: http://www.shmu.sk/sk/?page=21, accessed 25 November 2010 S-Plus (2000): Professional edition for Windows, release 2. (1999). MathSoft, Inc., Massachusetts Šmídová, A. (2006): Popelivka sibiřská (Ligularia sibirica) na Jestřebských slatinách. Severočeskou Přír., 38: 11–17. Šmídová, A. (2009): Studie faktorů ovlivňující dynamiku druhu Ligularia sibirica na Slovensku. Prír Nízkych Tatier, Banská Bystrica, 2: 55–60. Šmídová, A., Münzbergová, Z. & Plačková, I. (2011): Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae). Flora, 206: 151–157. Thompson, K. & Grime, J.P. (1979): Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. J. Ecol., 67: 893–921. Tomimatsu, H. & Massashi, O. (2010): Demographic response of plant populations to habitat fragmentation and temporal environmental variability. Oecologia, 162: 903–911. Vergeer, P.V., Rengelink, R., Copal, A. & Ouborg, N.J. (2003): The interacting effects of genetic variation, habitat quality and population size in performance of Succisa pratensis. J. Ecol., 91: 18–26.

45 Chapter I

Electronic Supplementary Material

Table S1: Relationship between Ellenberg indicator values of nitrogen availability, moisture, soil reaction and light. Significant r values are in bold, P < 0.05

nitrogen moisture reaction moisture -0.56 reaction 0.46 -0.32 light -0.79 0.41 -0.1

Table S2: Average number of viable seeds in the first and second year seed bank. See Table 1 for abbreviations of the study localities

No. of viable seeds in the seed bank Population 1 st year 2 nd year CZ N/1 0 0 CZ N/2 0 0 CZ N/2 0 0 CZ N/3 4.2 0.8 CZ N/4 3.2 0.6 CZ N/5 2 0.4 CZ C/1 1.8 0.8 CZ C/2 0 0 CZ C/3 0 0 CZ S/1 1.2 0.4 SK/SR 0 0 SK/P 0 0

46 Chapter I

Table S3: Average numbers of initiated, developed seeds and of proportion of damaged seeds in study years. See Table 1 for abbreviations of the study localities

No. of initiated seeds No. of developed seeds Proportion of damaged seeds Population 2005 2006 2007 2005 2006 2007 2005 2006 2007 CZ N/1 33.13 33.95 33.13 19.37 18.48 19.80 0.23 0.06 0.07 CZ N/2 29.67 27.85 27.83 17.75 17.43 16.82 0.19 0.01 0.18 CZ N/3 38.70 41.05 40.07 31.05 29.38 32.17 0.03 0.01 0.01 CZ N/4 37.72 34.57 34.42 26.40 21.95 25.12 0.08 0.06 0.15 CZ N/5 29.57 33.58 30.43 13.65 13.32 16.20 0.18 0 0.07 CZ C/1 30.48 29.83 31.30 14.00 19.33 23.43 0.12 0.08 0.08 CZ C/2 37.60 36.97 36.45 26.67 24.42 27.22 0.05 0.05 0.12 CZ C/3 32.39 34.68 35.17 22.59 25.47 26.80 0.18 0.09 0.09 CZ S/1 29.02 31.50 31.55 13.88 16.83 21.43 0.11 0.03 0.01 SK SR 27.08 26.08 26.30 13.22 11.12 9.48 0.45 0.38 0.57 SK P/1 37.18 32.43 36.87 20.53 16.92 17.25 0.05 0.08 0.08

47 Chapter I

= seed bank first year, 2 = seed 2 = bank = seed seed bank first year, study localities1 see Table second year, 3 = seedlings, 4 = juveniles, 5 = vegetative adults, 6 = fertile adults, 7 = dormant adults. For abbreviations of of abbreviations For adults. dormant 7 = adults, fertile = 6 adults, = vegetative 5 = juveniles, 4 = seedlings, 3 year, second Table S4: Transition matrices for each population and transition interval. n = number of individuals in given size category. 1 category. size given in individuals of number n = interval. transition and population each for matrices Transition S4: Table

48 Chapter I

49 Chapter I

Table S5: Population growth rate (λ) of single populations of Ligularia sibirica within each transition interval with 95% confidence interval (CI). See Table 1 for abbreviations of the study localities

2005 - 2006 2006 - 2007 2007 - 2008 Population λ CI λ CI λ CI

CZ N/1 1.74 1.36 - 1.98 1.39 0.98 - 1.58 1.62 1.37 - 1.81 CZ N/2 1.81 1.56 - 1.99 1.32 1.12 - 1.46 1.44 0.98 - 1.62 CZ N/3 0.92 0.87 - 0.96 0.95 0.90 - 0.98 0.98 0.95 - 1.01 CZ N/4 0.94 0.89 – 0.99 0.93 0.87 - 0.98 0.93 0.89 - 0.98 CZ N/5 0.86 0.80 - 0.93 0.83 0.76 - 0.90 0.78 0.71 - 0.87 CZ C/1 1.09 0.99 - 1.18 1.00 0.95 - 1.06 1.17 1.04 - 1.27 CZ C/2 1.06 0.93 - 1.17 1.12 0.91 - 1.26 0.98 0.88 - 1.06 CZ C/3 1.13 0.98 - 1.25 1.22 0.99 - 1.35 1.36 1.13 - 1.51 CZ S/1 0.99 0.93 - 1.05 0.99 0.95 - 1.03 1.03 0.98 - 1.08 SK SR 1.32 1.15 - 1.45 1.20 1.10 - 1.28 1.15 1.08 - 1.21 SK P/1 1.32 1.05 - 1.51 1.30 1.10 - 1.44 1.03 0.91 - 1.15

50 Chapter I

ignificantly different from those different ignificantly sibirica using life-table response response life-table using sibirica year, 3 = seedlings, 4 = juveniles, 5 = 5 = juveniles, = 4 seedlings, 3 = year, st and indicate that the contributions are s that the contributions and indicate st population growth rates between populations of L. populations between rates growth population ance values were determined using a permutation te ngle matrix elements to the observed variation in variation observed the to elements ngle matrix expected at random. Significance levels: * P < 0.05, **P < 0.01, *** P < 0.001. 1 = seed bank first year, 2 = seed bank second 2 second = seed bank year, 1 first *** bank P < < = seed * **P P 0.001. 0.01, < 0.05, levels: Significance at random. expected 1 see Table localities of study For abbreviations adults. 7 dormant = adults, 6 = fertile adults, vegetative Table S6: Contribution of si Table S6: Contribution experiments (LTREs). The signific experiments (LTREs). The

51 Chapter I

Fig. S1: Ellenberg indicator values of nitrogen availability. Values represent mean ± standard Error

52 Chapter I

Fig. S2: Elasticity of single life cycle transitions for Ligularia sibirica with 95% confidence intervals. The values were obtained using stochastic simulation. The plots within a figure are arranged in a form of a transition matrix. See Table 1 for abbreviations of the study localities

53

CHAPTER II

Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae)

Šmídová A., Münzbergová Z. & Plačková I. (2011) Flora 206: 151-157

Chapter II

Abstract Rare plant species can be divided into naturally, ‘old rare’ species and anthropogenically, ‘new rare’ species. Many recent studies explored genetic diversity of ‘new rare’ species. Less is, however, known about genetic diversity of ‘old rare’ species. We examined isozyme genetic variability of 20 populations of an ‘old rare’ plant species, Ligularia sibirica (Asteraceae) in the Czech and Slovak Republic. It is a long lived perennial herb with mixed-mating breeding system, widely distributed from East Asia to European Russia, with few isolated relict populations in the remaining part of Europe. The results showed high genetic diversity within populations (80.81%) and a low level of genetic differentiation (FST=0.179). Genetic distance between populations correlated significantly with geographic distance. There was also a significant positive correlation between genetic diversity and population size. This is probably caused by destruction of habitats in last centuries and subsequent decrease of population size. Patterns of genetic diversity suggest that the recent distribution is a result of stepwise postglacial migration and consequent natural fragmentation. We conclude that L. sibirica populations preserve high levels of genetic diversity and are not yet threatened by genetic factors. However, this may change if the changes in habitat conditions continue.

Keywords: Allozyme, genetic diversity, Ligularia sibirica, old rare species, population size, postglacial migration

Introduction Genetic consequences of habitat fragmentation are a topic of many recent studies (reviewed by Young et al. 1996; Leimu et al. 2006; Honnay & Jacquemyn 2007). It has been shown that small plant populations often face increased inbreeding depression and reduced genetic variation due to genetic drift (e.g. Young et al. 1996; Frankham et al. 2006). This often leads to decreased fitness of individuals as well as decreased ability to adapt to changes in environment (e.g. Booy et al. 2000; Leimu et al. 2006). Most of the studies on effects of habitat fragmentation are dealing with processes that occurred in the last decades or centuries due to human activities (Gonzales & Hamrick 2005) such as changes in land use (e.g. Stoate et al. 2001; Robinson & Sutherland 2002; Jacquemyn et al. 2003). These studies are mainly focusing on anthropogenically rare (‘new rare’) species, i.e. species that were formerly much more common in a particular area, and populations became smaller, less abundant, and more isolated because of human influence (sensu Huenneke 1991). Genetic diversity of such a ‘new rare’ species may still reflect genetic diversity in the past (Booy et al. 2000).

56 Chapter II

In contrast to recently declining ‘new rare’ species, genetic diversity of so-called ‘old rare’ species is expected to show long term consequences of habitat fragmentation (Reisch 2001; van Rossum & Prentice 2004). ‘Old rare’ species are species that are naturally rare in a specific area, occurring in small, isolated populations (sensu Huenneke 1991). To correctly understand the status of such a species, we need to investigate the history of their populations using evidence from population genetics as well as from biogeography, ecology and vegetation science (Pott 1995). ‘Old rare’ species include narrow endemics (Pegtel 1998) and relicts, such as glacial and postglacial relicts (Pegte, 1998; Rasmussen & Kollmann 2004). Studies of genetic diversity of endemic plant species have shown less than half of the genetic diversity of widespread species, low percentage of polymorphic loci and mean number of alleles per polymorphic locus (Hamrick & Godt 1989). In contrast, studies on European glacial relict species with arctic-alpine distribution range such as Biscutella laevigata (Dannemann 2000), Saxifraga aizoides (Lutz et al. 2000) and Saxifraga paniculata (Reisch 2001) have shown much higher genetic diversity, percentage of polymorphic loci and number of alleles than endemics or species with narrow distribution range (Hamrick & Godt 1989). The Central European relict plant species thus clearly represent a specific group of species with unique patterns of genetic diversity. The glacial relicts with arctic-alpine distribution range have received considerable attention in the past (e.g. Bauert et al. 1998; Dannemann 2000; Lutz et al. 2000; Reisch 2001). Another group of relict species, i.e. species with a large continuous distribution area from East Asia to the western part of Russia and with the western edge of distribution range in Europe, are still largely ignored. In Europe such species occur in few isolated localities (Meusel & Jäger 1992). To our knowledge, genetic variability of relict west range- peripheral populations was studied only in Angelica palustris (Dittbrenner et al. 2005), Iris aphylla (Wróblewska et al. 2003; Wróblewska 2008), Iris sibirica (Kostrakiewicz & Wróblewska 2008) and Stipa capillata (Krzakowa & Michalak 2007; Hensen et al. 2009). These studies have shown variable results. While Dittbrenner et al. (2005) found a similar level of genetic variability as revealed for most glacial relict species, Kostrakiewicz and Wróblewska (2008) found very low level of genetic variability, similar to endemic species. The results of the latter study may be related to the extensive clonal growth in Iris sibirica. In the present study, we use isozyme electrophoresis to investigate genetic variability and genetic structure of the relict plant species, Ligularia sibirica in the Czech and Slovak Republic. Ligularia sibirica has a wide continuous Eurosiberian distribution from East Asia to the European part of Russia. In central and western Europe this species has only few isolated localities (Meusel & Jäger 1992) and is listed among the species of European importance (European Communities, 1992). Due to recent habitat change, the species began to decline at some localities (Hendrych 2003). We can therefore expect that genetic diversity at these populations might be reduced as in the new rare species.

57 Chapter II

We asked the following questions: (1) What is the present-day isoenzyme variability of L. sibirica and its distribution within and among populations, and among geographical regions? (2) What is the population genetic differentiation? (3) Is there a correlation between population size and genetic variability? We aimed to interpret the results in relation to biology of L. sibirica and discuss how the present pattern of genetic variability is affected by the possible history of the populations.

Materials and methods

Study species Ligularia sibirica (L.) Cass. (Asteraceae) is a perennial hemicryptophyte 50 to 160 (210) cm high. It has ground rosettes of leaves and a short rhizome. During flowering from mid July to the end of August, L. sibirica creates 1 to 8 flowering stems with 20 (up to 55) flower heads per inflorescence on average. The flower heads are yellowish and bisexual. The species is entomogamous, pollinated mainly by bees (Slavík, 2004). The oval achenes (further referred to as seeds) are on average 2.162 mg heavy (Šmídová unpubl.). Their pappus forms a tuft of hair longer than the achenes (Hegi 1929; Slavík 2004). Seeds are either dispersed by wind over a short distance or by gravity. Terminal velocity of L. sibirica seeds varies between 0.86 – 1.18 ms-1 (Šmídová unpubl.). The species is diploid (2n = 60) (Liu 2004). Both sexual and vegetative reproduction occurs. It has a ‘phalanx’ type of clonal plant growth form which represents a tight-packed advancing front of ramets with slow radial spread (sensu Doust 1981). Ligularia sibirica is a plant species which prefers full sunlight, but it also grows under open tree canopy where it diminishes seed production (Kukk 2003). It occurs in a variety of wetlands, such as humid grasslands, alkaline mires and alluvial woodlands with the groundwater level at surface. The optimum of this species is in the alliances Caricion davallianae, Magnocaricion elatae and in mosaics of these communities (Procházka & Pivničková 1999). The species has a wide Eurosiberian distribution range. The main continuous distribution range is from East Asia to Southern Siberia and to the European part of Russia. In Europe, there are few disjunct populations in Estonia, Latvia, Poland, Hungary, Romania, Croatia, Bulgaria, the Slovak Republic, the Czech Republic, Austria, and France (Mattauch 1936; Meusel & Jäger 1992). Localities in these countries are rather distant from the continuous distribution range of the species, originated most probably in the early postglacial period and thus represent rare remnants of former more extensive distribution. Therefore, L. sibirica is considered to be a postglacial relict (Hendrych 2003). The species has currently 10 localities with more than 50 individuals in the Czech Republic and about 15 localities in the Slovak Republic. Ligularia sibirica is classified as a ‘critically endangered’ species in the Czech Republic (Procházka 2001) and as a ‘vulnerable’ species in the Slovak Republic (Feráková et al. 2001). For

58 Chapter II manipulation of this species and entering its localities, we obtained permission form the Ministries of Environment of the Czech and Slovak Republic. This species is also protected by EU Habitat Directive, Annex II. of the Council of European Communities (1992).

Study populations For our study we sampled all 10 populations in the Czech Republic and 10 selected populations in the Slovak Republic covering most of the populations containing more than 50 adult individuals (Fig. 1). Populations were considered as separate, when located at least 100 m apart from each other. The populations in the Czech Republic are located in three regions. The first one (containing 5 populations) is in northern Bohemia near the village Jestřebí (50°36’8” - 50°36’22” N, 14°36’51” - 14°37’21” E). The second one (containing 4 populations) is in central Bohemia near the village Rečkov (50°29’38” - 50°30’3” N, 14°51’56” - 14°54’40” E). The last population is in southern Bohemia in Pošumaví by a fishpond ‘Olšina’ (48°47’34” N, 14°6’7” E). Eight of the sampled populations in the Slovak Republic are in a large area in NP Slovenský raj (48°52’32” - 48°55’40” N, 20°13’50” - 20°21’44” E). One isolated population is in NP Nízke Tatry near the village Liptovská Teplička (48°58’2” N, 20°5’23” E) and one rather distant is in Salvatorské lúky approximately 25 km west of Prešov (49°2’45” N, 20°56’30” E). The distance between populations within regions varied from 100 m to approximately 10 km. The minimum distance between the Czech and Slovak populations is 410 km.

Fig.1: Geographic map of sampling localities of Ligularia sibirica populations (codes in Table 1) in the Czech and Slovak Republic

59 Chapter II

Habitat conditions in most of the study localities seem suitable for L. sibirica with exception of northern Bohemia region, the Czech Republic (populations CZ N/3, CZ N/4 and CZ N/5, Table 1). Here the fens had been fragmented and partly meliorated in the past (Hendrych 2003). Historical data about distribution of L. sibirica in the Czech and Slovak Republic are available since the early 19th century. They are not very detailed but they cover all main regions from which the species is currently known. L. sibirica was first mentioned from the Czech Republic in 1814 from northern Bohemia (Hendrych 2003). The localities in central Bohemia are known since 1843. The locality in southern Bohemia was found in 1984. The nativity of this locality was questioned, but no clear conclusion on this issue was drawn (Hendrych 2003; Slavík 2004). Populations in the Slovak Republic in the NP Slovenský raj are known since 1866. The isolated population in the Salvatorské lúky was first mentioned in 1853 (Domin 1940). The locality in the NP Nízke Tatry was recorded by botanists only in year 2000 (Turis 2000), but the species was present at the locality probably for at least 50 years (Turis unpubl).

Sampling and genetic analyses Allozyme variation was analyzed for a total of 400 L. sibirica individuals sampled from 20 populations in the Czech and Slovak Republic (Table 1). In each population 20 randomly selected individuals separated by at least 5 m from each other were sampled. We decided not to sample seedlings, because seedlings often have a higher genetic variation than adults. Mixing seedlings and adults in different proportions would lead to biased estimates of genetic diversity of the populations (Mandák et al. 2006; Dostálek et al. 2009). Leave tissue was kept on ice during transport to laboratory and stored in a refrigerator until extraction. Approximately 70 mg of leaf tissue was mechanically ground with Dowex-Cl (1-X8) and quartz sand and homogenized on ice in extraction buffer (0.1 M tris-HCl pH 8.0, 70 mM 2-mercaptoethanol, 26 mM sodium metabisulfite, 11 mM ascorbic acid, 4% (w/v) polyvinylpyrrolidone). The extraction and electrophoresis followed Kaplan et al. (2002) except for diaphorase (DIA, E.C. 1.6.99.3). The staining solution of diaphorase consisted of 100 ml tris-HCl buffer pH 8, 26 mg NADH, 10 mg MTT and 4 mg 2, 6-dichlorphenol-indophenol. All gels were incubated in the dark at 35°C until bands appeared. Afterwards, all gels were thoroughly rinsed in distilled water, dried between two cellophanes and stored. Staining recipes are available on request. The following 11 enzymes were analysed: ADH, AAT, DIA, EST, LAP, PGI, PGM, PRX, 6-PGDH, SHDH and SOD. From those we chose five enzyme systems, which reveal a clear pattern for genetic analyses. These five enzyme systems are: aspartate aminotransferase (AAT, E.C. 2.6.1.1), diaphorase (DIA, E.C. 1.6.99.3), leucyl aminopeptidase (LAP, E.C. 3.4.11.1), superoxide dismutase (SOD, E.C. 1.15.1.1) and 6-phosphogluconate dehydrogenase (6-PGDH, E.C. 1.1.1.44). A total of 11 loci were assayed for this species – AAT (2), DIA (3), LAP (1), SOD (3) and 6-PGDH (2).

60 Chapter II

)

IS

______F

O g to states (CZ and SK) regions H E 0.360 (0.269) 0.360 (0.269) 0.331 (0.318) -0.117 (0.119

ations are given in parentheses parentheses in given are ations . The. populations are divided accordin 64 24 0 0.294 0.318 -0.084 0.318 0.294 24 64 0 64 22 0 0.292 0.364 -0.245 0.364 0.292 22 64 0 65.91 (5.806) 65.91 (5.806) 24.55 (1.637) = fixation index. Standard devi IS Ligularia sibirica 72.73 26 0 0.373 0.370 0.008 0.370 0.373 26 72.73 0 54.55 24 0 0.292 0.277 0.052 0.277 0.292 24 54.55 0 P TA U H P U TA 143 63.64 24 0 0.283 0.273 0.033 0.273 0.283 24 63.64 143 0 0.022 0.250 0.256 23 63.64 000 0 P 3.125 72.73 27 0 0.320 0.370 -0.156 0.370 0.320 27 72.73 3.125 0 2.714 63.64 23 0 0.270 0.359 -0.332 0.359 0.270 23 63.64 2.714 0 -0.027 0.286 0.279 24 63.64 2.857 0 -0.059 0.314 0.296 24 54.55 3.167 0 2.857 63.64 25 0 0.314 0.291 0.073 0.291 0.314 25 63.64 2.857 -0.107 0.264 0.238 25 63.64 3.143 0 0 = mean number of alleles per polymorphic locus, P = percentage of polymorphic loci, TA = total number of alleles, alleles, of number = total TA loci, polymorphic of = percentage P locus, polymorphic per of alleles number = mean P 364 2.875 72.73 26 0 0.329 0.386 -0.175 0.386 0.329 26 72.73 2.875 364 0 = expected heterozygosity, F heterozygosity, = expected E r the studied populations of 2.250 (0.144) 2.934 (0.203) = observed heterozygosity, H heterozygosity, = observed O

2 2

n size and genetic variability fo variability n size and genetic e CZ N/4 323 63. 2.857 2.182 N/4 e CZ č ka SK NT/1 ~ 1500 2.273 3.000 63.64 25 0 0.240 0.282 -0.176 0.282 0.240 25 63.64 3.000 2.273 1500 0 ~ NT/1 ka SK č e CZ N/3 193 63. 2.429 2.000 N/3 e CZ č r CZ N/5 123 2.091 2.857 63.64 21 0 0.238 0.272 -0.140 0.272 0.238 123 21 63.64 2.857 2.091 N/5 0 r CZ ů ka CZ C/3 ~ 7 000 2.182 3.167 ní dv č č kov kov CZ C/1 ~ 10 000 2.545 3.125 č Louky u NV Louky u Kloko CZ C/2 ~ 10 000 2.454 Slovak Republic (SK) 3 3 (SK) Republic Valcha southern Bohemia (S) Slovak CZ C/4 85 2.182 3. Olšina CZ S/1 224 2.182 3. NP Slovenský raj (SR) (SR) raj NP Slovenský Malé Zajfy Vernár Biela Voda (NT) NP Nízke Tatry Prešov (P) SK SR/6 Means ~ 150 SK SR/8 SK SR/7 ~ 200 ~ 18000 2.091 2.273 2.182 3.000 63.64 25 0 0.295 0.382 -0.295 Pusté Pole Mokrá Vysoký Dobšiná Stratená dolina SK SR/1 ~ 28000 SK SR/2 SK SR/5 Tepli Liptovská SK SR/3 SK SR/4 ~ 300 ~ 1000 ~ 5000 2.182 lúky Salvatorské ~ 20000 2. 2.272 2.364 2.545 2.625 SK P/1 2.750 2.875 3.125 72.73 ~ 2000 72.73 72.73 2.182 72.73 24 25 26 28 0 0.331 0 0 0 0.329 0.291 0.359 0.386 0.418 0.341 0.404 -0.166 -0.272 -0.171 -0.128 Slune (C) central Bohemia Re Rákosina u Robe Czech Republic (CZ) 2 A 2 A (CZ) Estimated Table 1: Estimates of populatio U = numberuniqueH alleles, of Republic ______Population Abrev. ______Czech northern Bohemia (N) Baronský rybník Robe Louky u CZ N/2 243 2.272 ______Vojenské louky louky Vojenské CZ N/1 126 2.182 pop. size pop. (N, C,mean and P). S, SR, number NT A = of alleles perlocus, A

61 Chapter II

Data analyses

Standard measures of genetic diversity were calculated with the program ARLEQUIN (Arlequin 3.11; Excoffier & Schneider 2005) for each population. Calculated measures of genetic diversity are mean number of alleles per locus (A) and per polymorphic locus (AP), percentage of polymorphic loci (P), total number of alleles (TA) and number of unique alleles (U). We calculated mean observed (HO) and mean expected (HE) heterozygosity for all loci (Nei 1973) using POPGENE (version 1.32; Yeh & Boyle 1997). Variance components and their significance level for variation among groups, among populations within groups and within populations were conducted using AMOVA (Arlequin 3.11; Excoffier & Schneider 2005). Variance components were calculated separately for regions and the two republics (Czech and Slovak Republic).

In POPGENE (version 1.32; Yeh & Boyle 1997) we calculated summary F-statistics (FIS, FIT and FST). FST was calculated as genetic differentiation between each population compared with all other 19 pooled populations. We also calculated Wright’s fixation index (FIS) to determine deviations from Hardy-Weinberg expectations for each locus and population using the method of Wright (1978). Using χ2 test, we tested the significance of deviations from Hardy-Weinberg equilibrium at each polymorphic locus and population. We did not estimate level of gene flow (Nm) because studied populations do not fit the assumptions of equal number of individual per population and no spatial structure. Also the newly reduced populations have clearly not yet reached equilibrium between migration and genetic drift (Whitlock & McCauley 1999). We tested associations between log-transformed geographic distance and corresponding genetic distance between pairs of populations (FST) with a Mantel test (Mantel 1967) for the whole data set and then also for the two countries and three regions separately using ARLEQUIN (Arlequin 3.11; Excoffier & Schneider 2005). We tested the relationship between population size and different measures of genetic diversity using linear regression (STATISTICA version 7.0; StatSoft, Inc. 2004).

Results Nine of the 11 loci (81.81%) were polymorphic at least in one population. Only SOD-3 and DIA-3 were monomorphic across all populations. Overall 32 different alleles were detected in the polymorphic loci. Two alleles, LAP-a and 6PGDH-2-b, occurred only in the Czech Republic and three alleles LAP-e, LAP-f and SOD-1-b, occurred only in the Slovak Republic. The mean number of alleles per locus (A) was 2.250 (range: 2.000 - 2.545), alleles per polymorphic locus (AP) was 2.934 (range: 2.429 – 3.167). The percentage of polymorphic loci (P) was 65.91% (range: 54.55 – 72.73%). Total number of alleles (TA) was 24.55 and ranged from 21 to 28. The mean expected

62 Chapter II

population heterozygosity (HE) was 0.360 and ranged from 0.238 to 0.373. The mean observed heterozygosity (HO) was 0.331 and ranged from 0.264 to 0.418 (Table 1). Analysis of molecular variance (AMOVA) showed that 80.81% of molecular variance can be found within populations, 12.65% among populations within Czech and Slovak Republic and only 6.54% among the two republics (Table 2). The results of AMOVA for the study regions showed 82.02% variation within populations, 10.60% among populations within regions and 7.38% among regions. All proportions of variance were significantly different from zero (AMOVA, P < 0.001).

The overall fixation index, FIS, of individuals relative to their population was -0.116 and the mean overall inbreeding coefficient, FIT, was 0.084. Genetic differentiation between populations over all loci and populations, FST, was 0.179 (Table 3). Out of the total 145 tests for deviance of 9 loci/population from Hardy-Weinberg equilibrium, 66 (45.52%) were significant (P < 0.01). Of these, each 50% had positive FIS values, indicating a heterozygote deficit, and 50% had negative FIS values, indicating a homozygote deficit. These deviations were scattered across populations and loci with no obvious trends. Overall, eight out of nine polymorphic loci deviated significantly (P < 0.01) from Hardy-Weinberg equilibrium.

Of these, 75% had positive FIS values and remaining 25% had had negative FIS values.

Table 2: Results of the analyses of molecular variance (AMOVA) for 20 populations of Ligularia sibirica

Among populations Locus Among regions (%) Within populations (%) within regions (%)

Df 1 18 778

LAP 0.258 n.s. 18.989 *** 80.752 *** AAT-1 17.043 ** 9.748 *** 73.208 *** AAT-2 12.899 n.s. 15.573 *** 71.528 *** 6PGDH-1 12.318 n.s. 18.036 *** 69.646 *** 6PGDH-2 16.177 *** 11.039 *** 72.784 *** SOD-1 4.211 n.s. 8.596 *** 87.193 *** SOD-2 0.208 n.s. -1.544 ** 101.336 ** DIA-1 8.048 *** 2.739 *** 89.213 *** DIA-2 -1.114 n.s. 19.055 *** 82.057 ***

Overall 6.54*** 12.65*** 80.81*** Note: Overall values are results from overall AMOVA analyses, i.e. not averaged over single loci. The P values are derived using 1,000 permutations of the data. Significance levels * P < 0.01, **P < 0.005, *** P < 0.001, n.s. = not significant

63 Chapter II

Table 3: F statistics and average gene flow for Ligularia sibirica per polymorphic locus. (A) and over all loci (B) for each study region and two states (Czech and Slovak Republic). FIS = fixation index, FIT = inbreeding coefficient, FST = genetic differentiation between each population compared with all 19 other pooled populations

Locus FIS FIT FST (A) LAP 0.209 0.371 0.205 AAT-1 0.088 0.287 0.218 AAT-2 0.094 0.318 0.247 6PGDH-1 0.017 0.281 0.268 6PGDH-2 -0.083 0.159 0.224 SOD-1 -0.063 0.071 0.126 SOD-2 -0.965 -0.959 0.003 DIA-1 -0.800 -0.635 0.092 DIA-2 0.517 0.610 0.194

Overall -0.116 0.084 0.179

(B) Regions Northern Bohemia -0.096 0.010 0.097 Central Bohemia -0.017 0.075 0.091 Southern Bohemia 0.022 0.022 NA NP Slovenský raj -0.194 -0.034 0.134 NP Nízke Tatry -0.176 -0.176 NA Prešov -0.059 -0.059 NA

States Czech Republic -0.051 0.083 0.128 Slovak Republic -0.179 0.007 0.158

Most of the pairwise genetic distances (FST) between populations (189/190) were significant

(P < 0.05). The FST values ranged from 0.420 to -0.015. Pair wise genetic distances were significantly correlated with geographic distances (Mantel test; r = 0.395, P < 0.001) (Fig. 2). This correlation was also significant within the Czech Republic (Mantel test; r = 0.445, P = 0.002) and within the Slovak Republic (Mantel test; r = 0.485, P < 0.001) when calculated separately. However, when testing regions, a positive significant correlation was found only among populations in region NP Slovenský raj in the Slovak Republic (Mantel test; r = 0.715, P < 0.001) but not among regions northern Bohemia and central Bohemia in the Czech Republic (Mantel test; r = 0.392, P = 0.263 and r = 0.649, P = 0.163 respectively). There are significantly positive correlations of logarithm of estimated population size and measures of genetic diversity. Mean expected heterozygosity (HE; r = 0.561; P = 0.010) increased with population size (Fig 3). Similar results were found for mean number of alleles per locus (A; r = 0.631, P = 0.003), total number of alleles (TA; r = 0.600, P = 0.005) and mean observed heterozygosity (HO; r = 0.517; P = 0.02). The different response variables were partly correlated (Table 4).

64 Chapter II

Table 4: The relationship between different measures of genetic variation of Ligularia sibirica. A = mean number of alleles per locus, AP = mean number of alleles per polymorphic locus, P = percentage of polymorphic loci, TA = total number of alleles, HO = observed heterozygosity, HE = expected heterozygosity, FIS = fixation index. Significance values in bold, P < 0.05

A AP P TA HO HE FIS A

AP 0.54 P 0.64 -0.21 TA 0.91 0.47 0.57

HO 0.43 -0.32 0.67 0.47

HE 0.62 0.02 0.58 0.61 0.75

FIS 0.07 0.51 -0.32 0.02 -0.63 0.04

Discussion Main species traits that affect the amount of genetic diversity of plant species according to Hamrick and Godt (1989) are: life form, geographical range and breeding system. Long lived, perennial, widespread, outcrossing species have generally more genetic diversity than short lived, selfing or mixed-mating, narrow endemic species (Hamrick & Godt 1989, 1996). Ligularia sibirica is a long-lived, perennial herb with mixed-mating breeding system, in Europe classified as a relict. With this combination of species traits, genetic diversity of L. sibirica identified using allozymes should be somewhere in between low and high level, when compared with mean genetic diversity values based on allozymes given in Hamrick and Godt (1989). Observed values are on one hand close to endemic species (mean number of alleles per locus (A) = 3.00; L. sibirica: 2.250, SD ± 0.144), on the other hand roughly followed expectation for widespread species (percentage of polymorphic loci (P) = 59%; L. sibirica: 65.91%, SD ± 5.806 and mean expected heterozygosity

(HE) = 0.202; L. sibirica: 0.360, SD ± 0.269). A similar pattern was found also in studies of some other rare species based on allozymes, suggesting that the variation in genetic diversity between different rare species can be large (e.g. Dannemann 2000; Lutz et al. 2000). The analysis of molecular variance revealed most of the genetic variability (80.81%) within populations of L. sibirica, 12.65% variance among populations within regions and only 6.54% among regions. In case of other disjunctly distributed taxa the analysis showed similar high levels of intra-population diversity (Dannemann 2000; Lutz et al. 2000; Wróblewska et al. 2003; Dittbrenner et al. 2005). Such a higher genetic variability within, rather than among populations is typical for outcrossing, perennial plant species (Hamrick & Godt 1989, 1996). Thus, our findings may result from particular life history traits of L. sibirica, such as both sexual and vegetative reproduction, pollination by bees and long life span that tend to preserve genetic variability within populations (Loveless & Hamrick 1984).

65 Chapter II

0.45

0.40

0.35

0.30

0.25

ST 0.20 F

0.15

0.10

0.05

0.00

-0.05 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 log distance (km)

Fig. 2: The relationship between pairwise population FST values and logarithm of geographical distance (km). The relationship is significant (r = 0.395, P < 0.001, Mantel test). Dots around 2.6 log distances represent comparisons of Czech vs. Slovak populations

0.38

0.36

0.34

0.32

E 0.30 H

0.28

0.26

0.24

0.22 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Log population size

Fig. 3: The relationship between estimated population size and expected heterozygosity of Ligularia sibirica (r = 0.561; P = 0.01)

66 Chapter II

Genetic distance significantly correlated with the logarithmic geographic distance in the whole data set as well as in each country separately. This finding is in agreement with isolation by distance model (Ellstrand & Elam 1993) and corresponds with many other studies testing correlation by distance in natural plant populations (e.g. Reisch 2001; Hensen et al. 2005; Wróblewska 2008). This trend, however, was not observed for northern and central Bohemia when testing the regions separately. Observed isolation by distance may be generally caused by limited seed or pollen dispersal (Gaudeul et al. 2000). In our study, significant genetic differentiation was found even for populations that are 100 meters apart. This finding is in agreement with Ehrlich and Raven (1969), who observed that already distances of 15 metres to a few kilometres can effectively isolate two populations of insect pollinated plant species. Long distance dispersal of seeds is also quite unlikely. Even thought, L. sibirica seems to have good wind-dispersal seed potential (i.e. diaspores are achenes with pappus) and the terminal velocity is quite high (0.86-1.18 ms-1, Šmídová unpubl.). Within the seed dispersal values presented by Soons (2006) for wind dispersed wetland plants, L. sibirica belongs to species with maximal dispersal distances of several tens of meters. Additionally it has to be kept in mind that generally most seeds are dispersed close to the parental plant (Nathan & Muller-Landau 2000; Tackenberg et al. 2003). High within population genetic variability together with low genetic differentiation may suggest high gene flow. Alternatively, such pattern may be a result of past events (van Rossum & Prentice 2004). One hypothesis explaining this pattern in L. sibirica could be that distribution of genetic diversity within and between populations still reflects gene flow in the past before the formerly larger populations were fragmented (Hensen et al. 2005; Lowe et al. 2005; Peterson et al. 2008). The second hypothesis to be considered is that current gene flow connects populations that are geographically closer to each other more efficiently than populations divided by greater distances (Hensen et al. 2005; Peterson et al. 2008). This option is likely for northern and central Bohemia regions. We assume that both above proposed hypotheses might be true for L. sibirica. Similar results were found for other glacial relict species (e.g. Dannemann 2000; Lutz et al. 2000; Reisch et al. 2003). Populations of the glacial relicts were probably more widespread in Pleistocene (Reisch 2001) and have maintained their high genetic diversity in spite of approximately 12 000 years of isolation. Similarly to glacial relicts L. sibirica was probably more widespread in the past and its populations are presumably naturally fragmented for approximately 10 000 – 7 000 years. However, in literature we can find several theories on the origin of the species in Central and Western Europe. One, but today overcome idea is import of seeds with hay to feed the Cossack horses during the Napoleonic Wars from Russia (Kneblová 1950; Procházka & Pivničková 1999). Another hypothesis is a rare long-distance dispersal by migrating or overwintering birds (Walter & Straka 1970; Maršáková-Němejčková 1973). Such rare events would have to be repeated several times and several founder effects would lead to loss of alleles and to homozygosity (Hewitt 1999).

67 Chapter II

Our results thus contradict this hypothesis. The commonly accepted explanation is migration of L. sibirica sometime during the glacial period (Pleistocene, Mattauch, 1936) or in postglacial period (early Holocene, Meusel & Jäger 1992; Hendrych 2003). Most probably L. sibirica came to Europe from north-east in early Holocene, when the climatic conditions and vegetation in Europe were similar to conditions in the current centre of its distribution in southern central Siberia (Hendrych 2003) (high moisture and temperature in summer, low temperature in winter, Meusel & Jäger 1992). High level of genetic variability and low genetic differentiation observed in L. sibirica is characteristic for slower unidirectional stepwise migration (isolation by distance, Hewitt 1999; Gaudeul et al. 2000). We have no clear idea how widespread L. sibirica was in that time in Europe, but certainly, it was present in a much broader area than today. Consequently, most of the recent populations are remnants of former larger populations that persisted in climatically suitable locations with appropriate hydric conditions (Hendrych 2003). This hypothesis is further supported by the positive relationship between genetic and geographic distance. All the above suggests that L. sibirica is an ‘old rare’ species with the ability to maintain good level of genetic variability for a longer period of time. In this study we also found a significant effect of population size on genetic diversity measured as mean number of alleles per locus, total number of alleles, observed and expected heterozygosity, i.e. a pattern commonly reported to ‘new rare’ species (e.g. Fischer & Matthies 1998; Leimu et al. 2006 for reviews; but see Oostermeijer et al. 1994; Bachmann & Hensen 2007) and as well in ‘old rare species’ (Reisch et al. 2003; Dittbrenner et al. 2005). Thus our finding is in agreement with the hypothesis that small populations cannot maintain genetic diversity as high as those found in larger populations (Frankham et al. 2006). Such a feature is usually explained by genetic drift, by either random loss of alleles, founder effects during colonization and/or inbreeding events (Soulé 1986). Genetic drift, concretely random loss of alleles, may be true for some small sized populations of L. sibirica in northern Bohemia, the Czech Republic which have experienced relatively recent reduction of population size due to negative changes of habitat conditions in last centuries such as melioration and fragmentation of fens (Hendrych 2003). Founder effect may be explanation for one population (CZ S/1) which has been suggested that it may be of secondary origin by intentional planting (Hendrych 2003; Slavík 2004). We assume that if this was the case, the expected genetic diversity would be even lower than it would correspond to its size. This was, however, not the case and we thus do not have any support for this speculation.

Conclusion The results of our study on genetic diversity of an ‘old rare’ species, Ligularia sibirica showed high genetic diversity within populations and low level of genetic differentiation between populations which roughly follows values for outcrossing species. Genetic distance between populations correlated significantly with geographic distance suggesting the isolation by distance

68 Chapter II model for our study species. This model is in agreement with our hypothesis that L. sibirica migrated to Europe via slow unidirectional stepwise migration. There was also a significant positive correlation between genetic diversity and population size. This is probably caused by destruction of habitats of the species in fens in northern Bohemia, the Czech Republic, in last centuries and subsequent decrease of population size. Nevertheless, L. sibirica populations still preserve high levels of genetic diversity and are not yet threatened by genetic factors. However, this may change if the changes in habitat conditions continue.

Acknowledgments We would like to thank to workers of protected areas of the Czech and Slovak Republic in which L. sibirica occurs, for providing us with information about localities, J. Šmída for preparing the map and T. Heinken for help in field and useful comments on the manuscript. This study was supported by grant MŠMT 2B06178 and partly also by MŠMT 0021620828 and AV0Z60050516.

References Bachmann, U. & Hensen, I. (2007): Is declining Campanula glomerata threatened by genetic factors? Plant Species Biol., 22: 1-10. Bauert, M.R., Kälin, M., Baltisberger, M. & Edwards, P.J. (1998): No genetic variation detected within isolated relict populations of Saxifraga cernua in the Alps using RAPD markers. Mol. Ecol., 7: 1519-1527. Booy, G., Hendriks, R.J.J. Smulders, M.J.M., van Groenendael, J.M. & Vosman, B. (2000): Genetic diversity and the survival of populations. Plant Biol., 2: 379–395. Council of European Communities (1992): Council Directive 92/43/EEC of 21 May on the conservation of natural habitats and of wild fauna and flora. Off. J. Eur. Communities, 35: 7- 50. Dannemann, A. (2000): Der Einfluss von Fragmentierung und Populationsgröße auf die genetische Variation und Fitness von seltenen Pflanzenarten am Beispiel von Biscutella laevigata (Brassicaceae). Diss. Bot. 330. J.Cramer, Berlin, Stuttgart. (In German) Dittbrenner, A. Hensen, I. & Wesche, K. (2005): Genetic structure and random amplified polymorphic DNA diversity of the rapidly declining Angelica palustris (Apiaceae) in Eastern Germany in relation to population size and seed production. Plant Species Biol., 20: 191– 200. Domin, K. (1940): Vegetační obrazy ze Slovenska. Carpatica, Praha ser. B. 2: 51-72. Dostálek, T., Münzbergová, Z. & Plačková, I., (2010): Genetic diversity and its effect on fitness in an endangered plant species, Dracocephalum austriacum L. Conserv. Genet., 11: 773-783. Doust, L.L. (1981): Population dynamics and local specialization in a clonal perennial (Ranunculus repens): I. The dynamics of ramets in contrasting habitats. J. Ecol., 69: 743-755. Ehrlich, P.R. & Raven, P.H. (1969): Differentiation of populations. Science, 165: 1228-1232. Ellstrand, N.C. & Elam, D.R. (1993): Population genetic consequences of small population size: implications for plant conservation. Annu. Rev. Ecol. Evol. Syst., 24: 217-242. Excoffier, L.G.L. & Schneider, S. (2005): Arlequin ver. 3.0: An integrated software package for population genetics data analysis. Evol. Bioinformatics Online, 1: 47–50. Feráková, V., Maglocký, Š. & Marhold, K. (2001): Červený zoznam papraďorastov a semenných rastlín Slovenska (december 2001). In Baláž, D., Marhold, K. & Urban, P. (eds) Červený zoznam rastlín a živočíchov Slovenska. Ochr. Prír., Banská Bystrica, 20: 48–81. Fischer, M. & Matthies, D. (1998): RAPD variation in relation to population size and plant fitness in the rare Gentianella germanica (Gentianaceae). Am. J. Bot., 85: 811–819.

69 Chapter II

Frankham, R., Ballou, J.D. & Briscoe, D.A. (2006): Introduction to Conservation Genetics. Cambridge University Press, Cambridge Gaudeul, M., Taberlet, P. & Till-Bottraud, I. (2000): Genetic diversity in an endangered alpine plant, Eryngium alpinum L. (Apiaceae), inferred from amplified fragment length polymorphism markers. Mol. Ecol., 9: 1625-1637. Gonzales, E. & Hamrick, J.L. (2005): Distribution of genetic diversity among disjunct populations of the rare forest understory herb, Trillium reliquum. Heredity, 95: 306-314. Hamrick, J.L. & Godt, M.J.W. (1989): Allozyme diversity in plant species. In: Brown, A.H.D., Clegg, M.T., Kahler, A.L. & Weir, B.S., (eds) Plant population genetics, breeding and genetic resources. Sinauer, Sunderland, MA, pp 43–63. Hamrick, J.L. & Godt, M.J.W. (1996): Effects of life history traits on genetic diversity in plant species. Phil. Trans. R. Soc. Lond. B, 351: 1291-1298. Hegi, G. (1929): Illustrierte Flora von Mittel – Europa, Ed. 1. Vol. 6/2, München, pp 796-799. Hendrych, R. (2003): Poznatky o druhu Ligularia sibirica v Čechách. Preslia, 75: 39-69. Hensen, I., Oberprieler, C. & Wesche, K. (2005): Genetic structure, population size, and seed production of Pulsatilla vulgaris Mill. (Ranunculaceae) in central Germany. Flora, 200: 3– 14. Hensen, I., Kilian, C., Wagner, V., Durka, W., Pusch, J. & Wesche, K. (2010): Low genetic variability and strong differentiation among isolated populations of the rare steppe grass Stipa capillata L. in Central Europe. Plant Biol., 12: 526-536. Hewitt, G.M. (1999): Post-glacial re-colonization of European biota. Bot. J. Linn. Soc., 68: 87-112. Huenneke, L.F. (1991): Ecological implications of genetic variation in plant populations. In: Falk, D.A. & Holsinger, K.E. (eds) Genetics and conservation of rare plants. Oxford University Press, Oxford, pp 31-44. Honnay, O. & Jacquemyn, H. (2007): Susceptibility of common and rare plant species to the genetic consequences of habitat fragmentation. Conserv. Biol., 21: 823-831. Jacquemyn, H., van Rossum, F., Brys, R., Endels, P., Hermy, M., Triest, L. & de Blust, G. (2003): Effects of agricultural land use and fragmentation on genetics, demography and population persistence of the rare Primula vulgaris and implications for conservation. Belg. J. Bot., 136: 5-22. Kaplan, Z., Plačková, I. & Štěpánek, J. (2002): Potamogeton x fluitans (P. natans x P. lucens) in the Czech Republic II. Isozyme analysis. Preslia, 74: 187-195. Kneblová, V. (1950): Reservace u Rečkova a Ligularia sibirica (L.) Cass.. Čes. Bot. Listy, Praha, 3: 17-21. Kostrakiewicz, K. & Wróblewska, A. (2008): Low genetic variation in subpopulations on an endangered clonal plant Iris sibirica in southern Poland. Annu. Bot. Fennici, 45: 186-194. Krzakowa, M. & Michalak, M. (2007): Genetic variability of selected marginal populations of Stipa capillata L.. Biol. Lett., 44: 127-135. Kukk, Ü. (2003): The distribution of Ligularia sibirica (L.) Cass. in Estonia and changes in its population. Biul. Ogrodów Bot., 12: 11-22. Leimu, R., Mutikainen, P., Koricheva, J. & Fischer, M. (2006): How general are positive relationships between plant population size, fitness and genetic variation? J. Ecol., 94: 942– 952. Liu, J.-Q. (2004): Uniformity of karyotypes in Ligularia (Asteraceae: Senecioneae), a highly diversified genus of the eastern Qinghai–Tibet Plateau highlands and adjacent areas. Bot. J. Linn. Soc., 144: 329–342. Loveless, M.D. & Hamrick, J.L. (1984): Ecological determinants of genetic structure in plant populations. Annu. Rev. Ecol. Evol. Syst., 15: 65-95. Lowe, A., Harris, S. & Ashton, P. (2004): Ecological genetics: design, analysis, and application. Blackwell publishing, Oxford. Lutz, E., Schneller, J.J. & Holderegger, R. (2000): Understanding population history for conservation purposes: population genetic of Saxifraga aizoides (Saxifragaceae) in the lowlands and lower mountains north of the Alps. Am. J. Bot., 87: 583-590. Mandák, B., Bímová, K., Mahelka, V. & Plačková, I. (2006): How much genetic variation is stored in the seed bank? A study of Atriplex tatarica (Chenopodiaceae). Mol. Ecol., 15: 2653-2663.

70 Chapter II

Mantel, N. (1967): The detection of disease clustering and a generalized regression approach. Cancer Res., 27: 209-212. Maršáková-Němejčková M. (1973): Popelivka sibiřská. Ochr. Přírody, Praha, 28: 15–16. Mattauch, F. (1936): Über die Ligularia sibirica (L.) Cass.. Natur Heimat, Aussig 7: 57–62. Meusel, H. & Jäger J. (1992a, 1992b.): Vergleichende Chorologie der zentraleuropäischen Flora. Vol. 3. (1992a) Text, (1992b) Karten. G. Fischer, Jena. 491p. Nathan, R. & Muller-Landau, H.C. (2000): Spatial patterns of seed dispersal, their determinants and consequences for recruitment. Trends Ecol. Evol., 15: 278-285. Nei, M. (1973): Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci., 70: 3321-3323. Oostermeijer, J.G.B., Eijck, M.W. & Nijs, J.C.M. (1994): Offspring fitness in relation to population size and genetic variation in the rare perennial plant species Gentiana pneumonanthe (Gentianaceae). Oecologia, 97: 289–296. Pegtel, M.D. (1998): Rare vascular plant species at risk: recovery by seeding? App. Veg. Sci., 1: 67- 74. Peterson, A., Bartish, I.V. & Peterson, J. (2008): Effects of population size on genetic diversity, fitness and pollinator community composition in fragmented populations of Anthericum liliago L.. Plant Ecol., 198: 101-110. Pott, R. (1995): The origin of grassland plant species and grassland communities in Central Europe. Fitosociologia, 29: 7-32. Procházka, F. & Pivničková, M. (1999): Ligularia sibirica (L.) Cass. – In: Čeřovský, J., Feráková, V., Holub, J., Maglocký, Š. & Procházka, F. (1999): Červená kniha ohrozených a vzácnych druhov rastlín a živočíchov SR a ČR Vol. 5. Vyššie rostliny. Príroda a. s., Bratislava 219p. Procházka, F. (ed) 2001. Černý a červený seznam cévnatých rostlin České republiky (stav v roce 2000). Příroda, Praha 18: 1-166. Rasmussen, K.K. & Kollmann, J. (2004): Poor sexual reproduction on the distribution limit of the rare tree Sorbus torminalis. Acta Oecologica, 25: 211-218. Reisch, Ch. (2001): Climatic oscillations and the fragmentation of plant populations – genetic diversity within and among populations of the glacial relict plants Saxifraga paniculata (Saxifragaceae) and Sesleria albicans (Poaceae). Dissertation, University Regensburg. Reisch, Ch., Poschlod, P. & Wingender, R. (2003): Genetic variation of Saxifraga paniculata Mill. (Saxifragaceae): molecular evidence for glacial relict endemism in central Europe. Biol. J. Linn. Soc., 80: 11-21. Robinson, A.R. & Sutherland, W.J. (2002): Post-war changes in arable farming and biodiversity in Great Britain. J. App. Ecol., 39: 157-176. Slavík, B. (2004): Ligularia Cass. In: Slavík, B. & Štěpánková, J. (eds.) Květena České republiky Vol. 7. Academia, Praha, pp 306-309. Soons, M.B. (2006): Wind dispersal in freshwater wetlands: Knowledge for conservation and restoration. Appl. Veg. Sci., 9: 271-278. Soulé, M.E. (1986): Viable populations for conservation. Cambridge University Press, Cambridge. StatSoft, Inc. 2004. STATISTICA (data analysis software system), version 7. www.statsoft.com Stoate, C., Boatman, N.D., Borralho, R.J., Rio Carvalho, C., de Snoo, G.R. & Eden, P. (2001): Ecological impacts of arable intensification in Europe. J. Environ. Manag., 63: 337–365. Tackenberg, O., Poschlod, P. & Bonn, S. (2003): Assessment of wind dispersal potential in plant species. Ecol. Monogr., 73: 191–205. Turis, P. (2000): Charakteristika rašelinísk Národného parku Nízke Tatry. In: Stanová V. (ed.) Rašeliniská Slovenska. DAPHNE, Inštitút aplikovanej ekológie, Bratislava, pp. 73-74. Van Rossum, F. & Prentice, H.C. (2004): Structure of allozyme variation in Nordic Silene nutans (Caryophyllaceae): population size, geographical position and immigration history. Biol. J. Linn. Soc. 81: 357–371. Walter, H. & Straka, H. (1970): Arealkunde - Floristisch-historische Geobotanik. - Einrührung in die Phytologie III/2, Stuttgart. Whitlock, M.C. & McCauley, D.E. (1999): Indirect measures of gene flow and migration: FST ≠ 1/ (4Nm+1). Heredity, 82: 117-125. Wright, S. (1978): Evolution and the genetics of populations, Vol. IV. Variability within and among natural populations. University of Chicago Press, Chicago.

71 Chapter II

Wroblewska, A., Brzosko, E., Czarnecka, B. & Nowosielski, J. (2003): High levels of genetic diversity in populations of Iris aphylla L. (Iridaceae), an endangered species in Poland. Bot. J. Linn. Soc., 142: 65-72. Wróblewska, A. (2008): From the center to the margins of geographical range: molecular history of steppe plant Iris aphylla L. in Europe. Plant Syst. Evol., 272: 49-65. Yeh, F.C. & Boyle, T.J.B. (1997): Population genetic analysis of co-dominant and dominant markers and quantitative traits. Belg. J. Bot., 129, 157. Young, A., Boyle, T. & Brown, T. (1996): The population genetic consequences of habitat fragmentation for plants. Trends Ecol. Evol., 11: 413-418.

72

CHAPTER III

The habitat requirements of the endangered wetland species Ligularia sibirica

Heinken-Šmídová A. & Münzbergová Z. [manuscript]

Chapter III

Abstract Ligularia sibirica is an endangered wetland perennial plant species. To explain habitat requirement of the study species, we described plant species composition and investigated soil and leaf biomass chemistry at all its known localities in the Czech and Slovak Republic. We also analysed the relationship between habitat conditions and species composition of the localities. The results showed that L. sibirica has the ability to grow and persist in a wide range of plant communities with preference to the Caricion davallianae alliance or transitions between Caricion davallianae and Molinion caeruleae communities, i.e. sites with high level of ground water, neutral to alkaline pH, high content of calcium carbonate, phosphorus deficiency and low biomass production. In contrast to our expectations nutrient content in the soil was a better predictor of species composition of the sites as well as of individual performance of L. sibirica, than nutrient content in the biomass. Unlike many other fen species L. sibirica has the ability to grow and compete under various nutrient regimes. Our findings showed that adult individuals that persist in deteriorating habitats due to drainage or abandoned habitats are of large size, while individuals of L. sibirica which grow in base rich, nutrient poor habitats are of small size with low number of leaf rosettes, flower heads and seeds. Those results together with results of our previous study on population dynamics of L. sibirica indicate that individual performance is not a good prediction of performance of the whole population and that detailed demographic studies are necessary to assess population viability and to give advices for long-term protection of the species.

Keywords: Nutrient enrichment; Population size; Rare plant species; Size of single individuals; Vegetation

Introduction Wetlands, particularly low productive fen communities, are one of the most threatened and declining ecosystems in Europe (European Communities 2007). These fens have high species diversity (Stanová 2000; Grootjans et al. 2005) and represent refugia for many highly endangered plant species (Hájková et al. 2009). Fragmentation of wetlands and changes of their habitat quality are caused by both natural and human factors. While natural threats such as natural succession and drought have always been present (Shine & de Klemm 1999), human activities such as drainage for agriculture use, afforestation and land filling for urban development have strongly increased the rate of change during the last 2 000 years (Hartig et al. 1997), especially during the last two centuries (Shine & de Klemm 1999; European Communities 2007). Growth of plant species in these low productive wetlands is restricted by availability of nitrogen and on extreme calcium rich soils also by availability of phosphorus (Boeye et al. 1997; van Duren & Pegtel 2000). Species occurring in these habitats thus need to be adapted to this type of nutrient limitation (Wassen et al. 2005).

74 Chapter III

The main threats to survival of species in these habitats are connected with decrease in water table (van Duren & Pegtel 2000; Stanová 2000). Drainage of the habitats affects species not only directly due to limited availability of water for the plants, but mainly because by the drainage of peat, nitrate becomes available in larger amounts, whereas potassium availability may be reduced (van Duren et al. 1997; van Duren & Pegtel 2000). In addition, reduction or cessation of traditional management (Diemer et al. 2001; Peintinger & Bergamini 2006) leads to accumulation of biomass and litter in these habitats. All the changes in habitat conditions lead to changes in species composition and reduction of species diversity of both vascular plants (Pauli et al. 2002; Mälson et al. 2008) and bryophytes (Bergamini & Pauli 2001; Mälson et al. 2008). For example van Duren et al. (1997) or Pauli et al. (2002) who manipulatively added nutrients such as nitrogen (N), phosphorus (P) and potassium (K) to nutrient-poor fen localities, found that generalist plant species such as Filipendula ulmaria or Molinia caerulea increased aboveground biomass after N application, whereas the typical fen species such as Succisa pratensis or Potentilla palustris were suppressed under higher N, but positively reacted on P or K application. At first the change in vegetation composition is often visible as a change in fitness of individuals (e.g. number of leaves and rosettes, seed mass) and much later also as change in the population growth rate of the species (Brys et al. 2004) and in species composition (Pauli et al. 2002). Previous studies dealt with effects of changes in nutrient availability in nutrient-poor fens on composition of plant communities in these habitats (e.g. van Duren et al. 1997; Pauli et al. 2002). However, still little is known about the response of single typical fen species to these changes (but see Colling et al. 2002). To properly describe habitat conditions of the sites, it is necessary to get good estimates of nutrient availability. There are two possible ways to assess the availability of nutrients for plants: nutrient contents in the soil as basic measure, and in the biomass as widely accepted alternative method (e.g. Güsewell & Koerselman 2002; Rozbojová & Hájek 2008). In fens and bogs nutrient analyses of soils are mainly based on total contents of the upper peat layers, while exchangeable and thus available nutrients are seldom measured (Ellenberg & Leuschner 2010). In many soil types and especially for some elements such as nitrogen and phosphorus it is difficult to assess the nutrient availability. Firstly, what is extracted from the soil as available may not necessarily correspond to the chemical fractions actually available to plants (Verhoeven et al. 1988; Perez- Corona et al. 1996). Secondly, often even the concentration of available elements has limited explanatory power because the amount of available nutrients depends on their rate of mineralization (e.g. Rode 1995; Blume et al. 2010). Use of biomass nutrient concentrations may overcome the above mentioned problems of soil analyses and is assumed to integrate fluxes of nutrients to plants over a longer period of time (Güsewell & Koerselman 2002).

75 Chapter III

Here we study a plant species which mainly grows in nutrient-poor fen communities, Ligularia sibirica (L.) Cass (Asteraceae). It is a postglacial relict protected by the Habitats Directive, Annex II of the Council of the European Community (1992). Ligularia sibirica occurs mainly in nutrient-poor fen communities of the Caricion davallianae alliance, but preliminary data indicated that it also occurs in more productive communities. In this study we investigate the main environmental variables of habitats of the endangered perennial species Ligularia sibirica and their effect on plant morphological parameters and on population size. We also wanted to see if species composition of the localities and plant morphological parameters are related to nutrient content in the soil and in the biomass. Specifically, we asked the following questions. 1) What are the habitat requirements of the species in the Czech and Slovak Republic? 2) Which nutrient-related abiotic factors predict individual performance and population size of L. sibirica? 3) What is the importance of habitat conditions expressed as nutrient content in the soil and as nutrient content in the biomass for predicting species composition at the localities of L. sibirica and performance of L. sibirica? To describe habitat requirements of the species, we studied all 21 currently known populations of L. sibirica in the Czech and Slovak Republic with more than 50 adult individuals. To look at the relationship between habitat conditions and plant fitness, we followed the performance of L. sibirica in 13 of these populations.

Material and methods

Study species Ligularia sibirica (L.) Cass. (Asteraceae) is a perennial hemicryptophyte. It is 124.25 ± 43.00 cm high (mean ± SD) with ground rosettes of leaves and a short rhizome. During flowering from mid-July to the end of August, L. sibirica creates 1.91 ± 1.73 flowering stems (mean ± SD) with 18.79 ± 8.42 flower heads per inflorescence (mean ± SD). The flower heads contain 32.13 ± 7.26 achenes (mean ± SD) (Heinken-Šmídová A. & Münzbergová Z., unpublished data). Ligularia sibirica has hermaphrodite entomogamous flower heads (Slavík 2004). Sexual reproduction is highly predominant. Clonal growth of the ‘phalanx’ type, i.e. in the form of a tightly-packed advancing front of ramets with a slow radial spread (sensu Doust 1981), may also occur. The species prefers full sunlight, but it also grows under open tree canopy where seed production is lower (Kukk 2003). It occurs in a variety of wetlands, such as humid grasslands, calcareous fens, transitional mires and alluvial woodlands with the groundwater table at the surface (Slavík 2004). The species has a wide Eurosiberian distribution range. The main continuous distribution range is from East Asia to Southern Siberia and to the European part of Russia. In Europe, there are few disjunct populations in Estonia, Latvia, Poland, Hungary, Romania, Croatia, Bulgaria, the

76 Chapter III

Slovak Republic, the Czech Republic, Austria, and France (Mattauch 1936; Meusel & Jäger 1992). Localities in these countries are rather distant from the continuous distribution range of the species. Ligularia sibirica is regarded as a postglacial relict (Hendrych 2003; Šmídová et al. 2011), classified as a ‘critically endangered’ species in the Czech Republic (Procházka 2001) and as a ‘vulnerable’ species in the Slovak Republic (Feráková et al. 2001).

Study sites Data on vegetation composition of the sites, nutrient content in the soil and in biomass of L. sibirica were collected in the year 2006 from all known populations containing more than 50 adult individuals of L. sibirica in the Czech Republic and Slovak Republic (Fig. 1). The populations in the Czech Republic are located in three regions. The first one (containing five local populations) is in northern Bohemia near the village Jestřebí (50°36’8” - 50°36’22” N, 14°36’51” - 14°37’21” E). The second one (containing four local populations) is in central Bohemia near the village Rečkov (50°29’38” - 50°30’3” N, 14°51’56” - 14°54’40” E). The last population is in southern Bohemia in Pošumaví, by the fishpond ‘Olšina’ (48°47’34” N, 14°6’7” E). Eight of the sampled populations in the Slovak Republic are in a large area in NP Slovenský raj (48°52’32” - 48°55’40” N, 20°13’50” - 20°21’44” E). Two isolated population are in NP Nízke Tatry near the village Liptovská Teplička (48°58’2” N, 20°5’23” E) and near the village Telgárt (48°51’9” N, 20°11’38” E). The last, rather distant population is in Salvatorské lúky approximately 25 km west of Prešov (49°2’45” N, 20°56’30” E).

Vegetation description To describe habitat conditions at each locality, we recorded one to three phytosociological relevés in 5 × 5 m plots in each locality in such a way that the heterogeneity of the locality was covered. Species abundances of vascular plants and bryophytes were estimated by the Braun- Blanquet cover scale (see Moravec 1994).

Soil and biomass analysis We took 10 core soil samples (100 cm3 per core) from the rhizosphere (5-20 cm) at random from each locality and pooled them into a single sample. The samples were dried at 70 °C, sieved with a 2 mm mesh sieve and thoroughly mixed. We measured pH (H2O), content (%) of organic carbon (C) and content (%) of total nitrogen (N). Determinations of metallic cations (total content 2+ + of calcium (Ca ) and potassium (K )) and potentially available phosphorus concentrations (P2O5, furthermore denoted as ‘P’) in soil samples were made after extraction in a Mehlich II solution using a spectrophotometer and atomic absorption spectrometer. All procedures were made according to the standard methodology of Králová (1990) and Zbíral (1995).

77 Chapter III

For the biomass analysis we randomly selected 20 leaves from 20 sterile adult individuals of L. sibirica during the flowering time (July-August) from each locality. Leave samples of each locality were dried at 70 °C to constant weight, grounded and thoroughly mixed. We measured total content (%) of carbon (C) and nitrogen (N), calcium (Ca2+), potassium (K+) and phosphate (P) according to the standard methodology of Zbíral (1995).

Population size Size of L. sibirica populations was estimated in 2006. An adult individual of L. sibirica was defined as an individual consisting of 1 to 15 rosettes in a dense cluster that was connected by up to 10 cm long underground rhizomes most likely representing a single genet (Heinken-Šmídová A. & Münzbergová Z., in press). In the case of small populations (< 500 adult individuals), the population size was the sum of all adult individuals in the locality. In the case of large populations (> 500 adult individuals), the population size was estimated as an area of the population multiplied by the average number of adult individuals in ten 1 m2 plots. The plots were spread throughout the whole locality in order to cover the different densities of adult individuals in the proportions in which they occurred at the respective localities. Since higher errors can occur in the estimation of large populations, we used logarithmically transformed population size values for subsequent analyses.

Population characteristics In a subset of 13 populations of L. sibirica selected morphological characteristics were measured (Fig. 1). During the flowering time (mid July to end of August) in 2006 we recorded the number of leaves and the number of rosettes for each adult individual. For fertile plants the number and length of flowering stems, the length of inflorescence and the number of flower heads per inflorescence were additionally recorded. During the maturation of seeds (September) the number of initiated seeds per flower head on 20 randomly selected fertile plants were assessed. These data were used to estimate seed set per flower head and seed set per plant in each population.

Data analysis We assigned an Ellenberg indicator value for light, nitrogen availability, moisture and soil reaction (Ellenberg et al. 1992) to each plant species in each relevé if the value was available. Both vascular plant species and bryophytes were considered. The differences between relevés within a single locality were quite low, thus we used one representative phytosociological relevé at each locality and for those we thereafter calculated unweighted arithmetic means of Ellenberg indicator values. It should be noted that the Ellenberg N values indicate general nutrient availability rather than nitrogen availability (Schaffers & Sýkora 2000; Diekmann 2003; Ellenberg & Leuschner 2010). Therefore, in the text we will use Ellenberg N as ‘soil nutrient availability’ and increasing

78 Chapter III

Ellenberg N values will be referred to ‘eutrophication’. We used Ellenberg indicator values to describe habitat conditions, because they are widely and successfully used in this respect (Diekmann 2003; Käfer & Witte 2004). They were developed mainly on the basis of field experience and they reflect the ecological behaviour of species. Therefore, the average indicator values do not refer to habitat conditions at a specific moment, but present integration over time (Ellenberg et al. 1992; Schaffer & Sýkora 2000). Gradients in the composition of the vegetation of the sites were analysed by detrended correspondence analyses (DCA, Ter Braak & Šmilauer 2002). We used percentage species data with logarithmic transformation and rare species were down-weighted. The scores of vegetation relevés at the first DCA axis were included in the set of environmental variables and they represent habitat quality, i.e. the gradient from low habitat quality (degraded nutrient-rich sites) to high habitat quality (well preserved nutrient-poor sites). The analyses were performed in Canoco 4.5 (Ter Braak & Šmilauer 2002).

Fig.1: Geographic map of sampling localities of Ligularia sibirica populations in the Czech and Slovak Republic. Localities of L. sibirica where the morphological characteristics where measured are underlined

List of localities and their abbreviations: Czech Republic (CZ) - Vojenské louky CZ N/1; Baronský rybník CZ N/2; Louky u Robeče CZ N/3; Rákosina u Robeče CZ N/4; Sluneční dvůr CZ N/5; Rečkov CZ C/1; Louky u NV CZ C/2; Klokočka CZ C/3; Valcha CZ C/4; Olšina CZ S/1. Slovak Republic (SK) - Pusté Pole SK SR/1; Mokrá SK SR/2; Vysoký SK SR/3; Dobšiná SK SR/4; Stratená dolina SK SR/5; Malé Zajfy SK SR/6; Vernár SK SR/7; Biela Voda SK SR/8; Liptovská Teplička SK NT/1; Telgárt SK NT/2; Salvatorské lúky SK P/1

79 Chapter III

The following parameters were used as environmental variables for statistical analyses: mean Ellenberg indicator values for light (Ellenberg L), nitrogen/nutrient availability (Ellenberg N), moisture (Ellenberg M) and soil reaction (Ellenberg R), content of C, N, Ca, K, P, C:N ratio in the soil and in the leaf biomass of L. sibirica, pH values in the soil and scores of the vegetation samples on the first DCA axis. We conducted two types of analyses. The first one was performed on the whole data set containing 21 populations. Canonical correspondence analysis (CCA) was used to evaluate the relationship between vegetation composition of the sites and the content of nutrients in the soil and in the leaf biomass. Significance was assessed using a Monte-Carlo permutation test available in Canoco (Ter Braak & Šmilauer 2002). We also used the vegetation relevés to assign the localities to specific vegetation types. The second analysis was performed on the subset of 13 populations for which selected morphological parameters of L. sibirica were recorded. Morphological parameters were number of leaves of vegetative and flowering individuals, length of inflorescence and number of seeds per individual was not used in the following analysis due to positive correlation with other morphological parameters. To assess the effect of all above mentioned environmental variables on morphological parameters, we first performed regression analyses for each dependent and independent variables separately. Specifically, we used generalized linear models with Gaussian distribution in the case of the logarithm of population size and the length of flowering stems and generalized linear models with Poisson distribution in the case of the number of rosettes of vegetative and fertile individuals, the number of flowering stems, the number of flower heads per inflorescence and the seed set per flower. In the next step we performed stepwise linear regression for each dependent variable. All statistical analyses were carried out in S-Plus (2000). Correlation analyses were performed between various dependent and independent variables (Table S1-4 in Electronic Supplementary Material (hereafter ESM)).

Results

Habitat requirements of Ligularia sibirica Based on phytosociological relevés, L. sibirica grows in a relatively wide range of plant communities. Because the habitats of the species strongly differ between the Czech and the Slovak Republic, they are described separately (Fig. 2). In Slovakia, most populations of the studied species grow in calcium-rich fens of the Caricion davallianae Klika 1934 alliance which belong to the Valeriano simplicifoliae-Caricetum davallianae Moravec 1966 association. This association contains species such as Calamagrostis varia, Cirsium rivulare, Primula farinosa, Swertia perennis and Trollius europaeus (Fig. 2). Ligularia sibirica also occurs in wet meadows of the Calthion palustris Tüxen 1937 alliance and

80 Chapter III mountainous alluvial forests of the Alnion incanae Pawłovski et al. 1928 alliance, on sites with the ground water level by surface. In Slovakia, the studied species occurs in the mountain ranges. The lowest occurrence is in the locality Salvátorské lúky approximately in 515 m a. s. l. and the highest occurrence is in the locality Telgárt approximately in 985 m a. s. l.. In this country L. sibirica also occurs in a wide range of soil and leaf biomass characters. The content of C in the soil ranges from 13.91% in SK SR/5 to 49.70% in SK NT/1, the content of N from 0.85% in SK SR/6 to 2.91% in SK SR/4, the content of Ca ranges from 1587 mg/kg in SK NT/1 to 22661 mg/kg in SK SR/8, the content of K from 28.9 mg/kg in SK P/1 to 339.4 mg/kg in SK SR/3, the content of P from 6.73 mg/kg in SK SR/6 to 30.11 mg/kg in SK SR/3 and the range of pH values are from 5.81 in SK NT/1 to 7.8 in SK SR/8. The content of C leaf biomass ranges from 42.71% in SK SR/5 to 44.39% in SK SR/8, the content of N from 1.44% in SK NT/1 to 2.94% in SK SR/5, the content of Ca ranges from 6487 in SK SR/1 to 10363 mg/kg in SK SR/4, the content of K from 21157 mg/kg in SK SR/4 to 34436 mg/kg in SK SR/8 and the content of P from 817.07 mg/kg in SK SR/8 to 1951 mg/kg in SK SR/5. Most of the localities of L. sibirica in Slovakia were formerly pastured, grazed or used for hay- making, but today they are not longer managed and thus threatened by succession to Calthion communities (Stanová 2000). In the Czech Republic, L. sibirica mostly grows in the transition between the Valeriano dioicae-Caricetum davallianae Kuhn 1937 or Juncetum subnodulosi Koch 1926 associations within the Caricion davallianae alliance and seasonally wet meadows of the Molinion caeruleae Koch 1926 alliance. The above mentioned species of the Caricion davallianae in Slovakia do not occur here. Instead, species such as Cirsium palustre, Lysimachia vulgaris and Molinia caerulea are frequent (Fig. 2). However, some of the study sites can not be included in those communities. This is true for the location in Pošumaví in South Bohemia (CZ S/1), which is similar to the nutrient-poor fens, but less calcium-rich and thus more acidic. This location can be classified as transitional mire and a mosaic of the alliances Caricion fuscae Koch 1926, Molinion caeruleae and Calthion palustris. The second exception is the locality CZ C/2 in the alluvial plain of the river Rokytka in central Bohemia which can be classified as nutrient-rich sedge vegetation of the Magnocaricion elatae Koch 1926 alliance with Carex acutiformis. In all localities in the region of central Bohemia, L. sibirica occurs also in alder swamp forests of the Alnion glutinosae Malcuit 1929 alliance. The last exceptions are the localities CZ N/3 – 5 in Northern Bohemia. They were in the past repeatedly drained. Today, these habitats are difficult to assign phytosociologically and can be described as atypical, nitrogen and/or nutrient rich Molinietalia communities dominated by tall plants, such as Molinia caerulea and Phragmites australis (Fig. 2). In the Czech Republic, the studied species occurs mainly in lowlands between 215 – 260 m a. s. l. with the exception of the locality CZ S/1 in Pošumaví. This locality is approximately 730 m a. s. l. in a mountain range. In this country L. sibirica also occurs in a wide range of soil and leaf

81 Chapter III

Fig. 2: Species - samples biplot from DCA summarizing the relationship between species and samples. The species with the greatest weights were selected for display

Species list: Aln glu – Alnus glutinosa, Ang syl = Angelica sylvestris, Aul pal = Aulacomnium palustre, Cal var = Calamagrostis varia, Cam ste = Campyllium stellatum, Car acu = Carex acutiformis, Car dav = Carex davalliana, Cir ole = Cirsium oleraceum, Cir riv = Cirsium rivulare, Cli den = Climacium dendroides, Cre pal = Crepis paludosa, Des ces = Deschampsia cespitosa, Equ pal = Equisetum palustre, Fes rub = Festuca rubra, Fis adi = Fissidens adianthoides, Gal uli = Galium uliginosum, Pot ere = Potentilla erecta, Sph sub = Shagnum subnitens, Suc pra = Succisa pratensis, Swe per = Swertia perennis, Tro eur = Trollius europaeus, Val dio = Valeriana dioica

Samples list: ▲ Slovak Republic, Δ Czech Republic. See Fig. 1 for detail abbreviations of the study localities

82 Chapter III biomass characters: The content of C in the soil ranges from 4.38% in CZ C/1 to 22.710% in CZ N/2, the content of N from 0.40% in CZ C/1 to 3.92% in CZ N/5, the content of Ca ranges from 4147 mg/kg in CZ C/1 to 14439 mg/kg in CZ N/2, the content of K from 22.6 mg/kg in CZ C/4 to 479.6 mg/kg in CZ N/5, the content of P from 3.01 mg/kg in CZ N/3 to 18.06 mg/kg in CZ N/4 and the range of pH values is from 5.05 in CZ N/5 to 6.57 in CZ C/1. The content of C leaf biomass ranges from 40.35% in CZ C/4 to 44% in CZ N/5, the content of N from 1.52% in CZ C/1 to 2.94% in CZ N/4, the content of Ca ranges from 8971 in CZ C/4 to 16918 mg/kg in CZ C/2, the content of K from 12539 mg/kg in CZ N/3 to 28076 mg/kg in CZ N/2 and the content of P from 717.07 mg/kg in CZ N/5 to 1442 mg/kg in CZ N/2. Today, the habitats of L. sibirica which are or will be pronounced as protected (i.e. CZ C/1 and CZ C/3) are managed.

Relationship between vegetation data and the content of nutrients in the soil and in the leaf biomass In the 21 relevés 198 species (160 vascular plants and 38 bryophytes) were found. When looking at the relationship between vegetation data and the soil characters, the variation of vegetation composition was significantly influenced by pH values (% explained variability = 9.98%, P = 0.002), the content of C in the soil (% explained variability = 7.08 %, P = 0.02) and the content of N in the soil (% explained variability = 6.60 %, P = 0.03) (Fig. 3). The first canonical axis explained 11.93 % of the variation and was closely associated with the soil chemistry gradient pH/calcium, whereas axis two explained 6.49% of the variation. The pH/calcium gradient of the first canonical axis was a gradient from vegetation of base-rich substrates (high pH values and high content of calcium) with Carex davalliana, Crepis paludosa, Swertia perennis, Trollius europaeus, Fissidens adianthoides and Plagiomnium elatum and to vegetation of more base-poor substrates (lower pH values and lower content of calcium) with Alnus glutinosa, Lysimachia vulgaris, Molinia caerulea and Phragmites australis. The three significant values of nutrient content in the soil (pH, C and N) explained altogether 23.6 % of the variation in species composition of the sites. When looking at the relationship between vegetation data and the content of nutrients in the leaf biomass, the variation of vegetation composition was significantly influenced by the content of Ca (% explained variability = 9.74 %, P = 0.004), N (% explained variability = 7.83 %, P = 0.006) and P in the leaf biomass (% explained variability = 6.12 %, P = 0.04) (Fig. 4). The first canonical axis explained 13.71% of the variation and was closely associated with the nitrogen gradient (from low to high concentration of nitrogen in the leaf biomass), whereas axis two explained 5.37 % of the variation. The three significant values of nutrient content in the soil (Ca, N and P) explained altogether 23.7 % of the variation.

83 Chapter III

Fig. 3: Species - environment biplot from CCA summarizing the relationship between species and nutrient content in the soil. Significance above each variable indicates the result of a Monte-Carlo permutation test. Significance levels: **P < 0.01, * P < 0.05, n.s. - non significant. The species with the greatest weights were selected for display

Species list: Aln glu – Alnus glutinosa, Ang syl = Angelica sylvestris, Aul pal = Aulacomnium palustre, Cal cus = Calliergonella cuspidata, Cal var = Calamagrostis varia, Cam ste = Campyllium stellatum, Car acu = Carex acutiformis, Car dav = Carex davalliana, Car pan1 = Carex panicea, Car pan2 = Carex paniculata, Cir ole = Cirsium oleraceum, Cir riv = Cirsium rivulare, Cli den = Climacium dendroides, Cre pal = Crepis paludosa, Des ces = Deschampsia cespitosa, Equ pal = Equisetum palustre, Fes rub = Festuca rubra, Fil ulm = Filipendula ulmaria, Fis adi = Fissidens adianthoides, Gal uli = Galium uliginosum, Geu riv = Geum rivale, Lat pra = Lathyrus pratensis, Lig sib = Ligularia sibirica, Lys vul = Lysimachia vulgaris, Mol cae = Molinia caerulea, Phr aus = Phragmites australis, Pic abi = Picea abies, Pla ela = Plagiomnium elatum, Pot ere = Potentilla erecta, Sph sub = Shagnum subnitens, Suc pra = Succisa pratensis, Swe per = Swertia perennis, Tro eur = Trollius europaeus, Val dio = Valeriana dioica

84 Chapter III

Fig. 4: Species - environment biplot from CCA summarizing the relationship between species and nutrient content in the leaf biomass. Significance above each variable indicates the result of a Monte- Carlo permutation test. Significance levels: **P < 0.01, * P < 0.05, n.s. - non significant. The species with the greatest weights were selected for display

Species list: Aln glu – Alnus glutinosa, Ang syl = Angelica sylvestris, Aul pal = Aulacomnium palustre, Cal cus = Calliergonella cuspidata, Cam ste = Campyllium stellatum, Car dav = Carex davalliana, Car pan1 = Carex panicea, Car pan2 = Carex paniculata, Cir ole = Cirsium oleraceum, Cli den = Climacium dendroides, Cre pal = Crepis paludosa, Des ces = Deschampsia cespitosa, Equ pal = Equisetum palustre, Fes rub = Festuca rubra, Fil ulm = Filipendula ulmaria, Fis adi = Fissidens adianthoides, Gal uli = Galium uliginosum, Lat pra = Lathyrus pratensis, Lig sib = Ligularia sibirica, Lys vul = Lysimachia vulgaris, Mol cae = Molinia caerulea, Phr aus = Phragmites australis, Pic abi = Picea abies, Pla ela = Plagiomnium elatum, Pot ere = Potentilla erecta, Sph sub = Shagnum subnitens, Suc pra = Succisa pratensis, Swe per = Swertia perennis, Tro eur = Trollius europaeus, Val dio = Valeriana dioica

85 Chapter III

When looking at the relationship between vegetation data and the content of nutrients in the soil and in the leaf biomass in one analysis, the variation of vegetation composition was significantly influenced only by the pH values and by the content of nutrients in the soil as described above.

Effects of environmental variables on a set of morphological measurements Of all the 18 recorded environmental variables, 14 were significantly correlated with at least one of the studied population morphological variables or with population size (Table 1). The morphological variables were mainly positively correlated with environmental variables such as Ellenberg N, content of K in the soil, content of Ca, C and N in the biomass. Further, they were negatively correlated with habitat quality (scores of the first DCA axis), Ellenberg L, content of Ca and C in the soil, pH, C:N ratio in the soil and in the biomass. Most of the correlations are for the number of rosettes of both vegetative and fertile individuals, the length of flowering stems and the number of seeds per flower head, while the number of flowering stems is rarely correlated with environmental variables. When looking at the environmental variables, the habitat quality (scores of the first DCA axis), Ellenberg L and Ellenberg N, C:N ratio in the soil and pH were the best predictors for morphological performance of L. sibirica, while Ellenberg R and M, content of N, P, K in the soil and content of K, P and C in the biomass showed no or very low relationship to the morphological variables. Specifically, we can say that the size of L. sibirica individuals, represented by the number of rosettes, the length of flowering stems and the number of seeds per flower head, increase with higher nutrients (i.e. Ellenberg N), content of Ca and N in biomass. On the other hand, the individuals are smaller under high amount of light (Ellenberg L), in high quality habitats (Fig. 5), high content of Ca and C in the soil, high C:N ratio in the soil (Fig. 6) and biomass and in low pH (Table 1). The results of regression analyses between environmental variables and population size showed a different pattern. The size of the populations significantly increases with increasing pH, content of P in the biomass, C:N ratio in the soil and biomass. On the other hand, the size of the populations significantly decreases with increasing content of K in the soil, N in the soil and leaf biomass (Table 1). When testing the effect of all environmental variables together using stepwise analyses, only very few independent variables were retained in the model due to their high correlation. Specifically, habitat quality significantly explained variation in the number of rosettes of vegetative individuals. The C:N ratio in the soil significantly explained variation in the number of rosettes of fertile individuals and in the length of flowering stems. The content of C in the soil also significantly explained variation in the length of flowering stems and in the number of flower heads

86 Chapter III

** ** * * * * marginally + surements and surements 50.72 62.39*** 45.88 36.73 44.89 31.88 38.00 — ↑ ↑ — ↑ — ↓ — ↓ — ↑ ↓ ses: —

** ** +

47.05 48.09 63.22*** 40.60* 23.50 32.29* 55.01** ↓ ↓ — ↓ ↓ — ↑ — ↓ — ↑ —

* *

32.66** 32.66** 17.49 25.01 ↓ ↑ ↓ — — — — — — — — — * * * * * * * * significantly negative. Df = 1, Df12. Errorshow The values = ↓

41.64 32.29 49.20** 49.20** 39.03 37.49 36.75 33.00 41.03 47.79** 47.79** 32.33 owering owering per heads per flower ↓ ↑ ↓ ↓ ↑ — ↓ — — ↑ ↓ ↓ ↓ selected in a step wise linear regression are in bold regression wise linear selected in a step

* le was tested separately. Significance levels of regression analy of regression levels Significance separately. le was tested

34.87* 34.87* 26.90 — — — — — ↓ ↓ significantly positive and and positive significantly ↑ — — —

— — * * * * + *

34.54 30.16 30.92 23.40 29.39 42.51 47.04** 47.04** ↓ ↑ ↓ ↓ — ↓ ↓ ↓ < 0.001. Environmental variables Environmental 0.001. < P

— — — — — * * * * * * * *

52.18** 52.18** 27.31 38.86 27.06 30.42 25.96 27.05 41.70 26.81 . Direction of the effect is the effect of . Direction ↓ ↑ ↓ ↑ ↓ ↑ ↑ ↓ ↑ < 0.01, *** P L. sibirica < 0.05, ** < 0.05, P head axis) (first DCA inflorescence stems stems Vegetative Fertile individuals No. of rosettes No. of rosettes flowering No. of Seed fl Length of Population No. of flower No. of seeds C biomass C biomass Ellenberg - L - L Ellenberg Ellenberg – N Ellenberg pH soil N biomass C:N biomass C:N biomass — — Ellenberg – M — — — — — M — — R – – Ellenberg Ellenberg soil C soil N soil — — — Ca soil K soil P C:N soil — — — biomass — — — — biomass — — K P Ca biomass Ca biomass individuals production size (log) (log) size individuals production ______Habitat quality ______significant, * Table1: Regression analyses of environmental variables (vegetation, soil and biomass composition) on a set of morphological mea morphological of a set on composition) biomass and soil (vegetation, variables environmental of analyses Regression Table1: variab independent of each in case deviance % of the explained population size size of population ______

87 Chapter III

Fig. 5: Relationship between habitat quality (the scores of vegetation relevés from the localities on the first DCA axis) and Length of flowering stems of Ligularia sibirica. See Fig. 1 for abbreviations of the study localities

Fig. 6: Relationship between C:N ratio in the soil and Number of rosettes of fertile individuals of Ligularia sibirica. See Fig. 1 for abbreviations of the study localities

88 Chapter III per inflorescence. The pH values significantly explained variation in the number of flowering stems and population size. The Ellenberg N significantly explained variation in the number of seeds per flower head. The variation in the number of seeds per flower head were also explained by C:N ratio in the soil (Table 1; values in bold).

Discussion

Habitat requirements of Ligularia sibirica Ligularia sibirica is a plant species which in Europe occurs in a variety of wetland habitats such as minerotrophic fens, wet and paludified grasslands and tall forb vegetation, willow scrubs, alluvial and spring forests. Minerotrophic fens of the Scheuchzerio-Caricetea fuscae Tüxen 1937 class are the most common habitat of L. sibirica. Those fens are reported as habitats of L. sibirica for example from Poland (Olaczek 2004), Estonia (Kukk 2003), Romania (Operea & Culiţă 2010) and France (Thebaud et al. 2007). Another habitat with frequent occurrence of L. sibirica is wet grasslands and tall forb vegetation of the Molinietalia Koch 1926 order and is reported as habitats of L. sibirica for example from Poland (Brơż & Przemyski 1988; Olaczek 2004), Estonia (Kukk 2003), Ukraine (Kobiv 2005) and France (Thebaud et al. 2007). Marginally, L. sibirica also occurs in wet scrubs or forests of the Salicion cinereae Müller et Görs ex Passarge 1961, Alnion glutinosae Malcuit 1929 and Alnion incanae Pawłovski et al. 1928 alliances. Those communities with presence of L. sibirica are reported from all countries of its occurrence. All this indicates that L. sibirica is able to grow or persist in a wide range of communities, soil pH/calcium and nutrient contents, light gradients and water regimes. In our study countries, the Czech and Slovak Republic, L. sibirica occurs preferably in nutrient-poor communities of the Caricion davallianae alliance or in the transition between Caricion davallianae and Molinion caeruleae alliances. These communities are characterized by a high level of ground or flowing water, high content of calcium carbonate, phosphorus deficiency, neutral to alkaline pH and a low biomass production (high C:N ratio in soil) (Rozbrojová & Hájek 2008). In those habitats we can find viable populations of L. sibirica (Heinken-Šmídová & Münzbergová, in press). Another common vegetation type of L. sibirica is wet and paludified grasslands and tall forb vegetation of the Molinietalia order. These are habitats of the Calthion alliance which developed from Caricion davallianae communities mainly after cessation of traditional use (Stanová 2000; Hájek & Hájková 2011), or meadow communities on nutrient-rich sites with dominance of Molinia caerulea and Phragmites australis which developed from drained former fens. The habitat conditions of nutrient-rich habitats are characterized by a long-term low level of ground water and due to the mineralization of peat also by released nutrients (Stanová 2000). Populations of L. sibirica today present in the above described altered type of Molinietalia communities are

89 Chapter III populations which highly probably existed there before habitat conditions changed. Those populations are in long perspective not viable. They are composed primarily from old adult individuals with very few or no seedlings and young individuals (Heinken-Šmídová & Münzbergová, in press). The study species also occurs in wetland forest communities such as Alnion glutinosae and Alnion incanae, especially if they are influenced by spring water. Those occurrences are mainly concentrated in the margins of those forests, because in shade the flowering and fruiting cease (Kukk 2003). Exceptionally, L. sibirica also occurs in nutrient-rich sedge vegetation of the Magnocaricion elatae alliance with ground water close to the surface during the whole year. In conclusion, the habitat range of L. sibirica in the Czech and Slovak Republic is similar to that reported from other European countries. In our study area calcareous fen communities prevail, while more acidic fen communities which are important for example in Estonia (Kukk 2003) are almost missing. Also in other countries populations of L. sibirica in successional or human altered vegetation such as Calthion or former Molinietalia communities with actual dominance of Phragmites australis (Stančić et al. 2010) or ruderal species such as Urtica dioica and Rubus idaeus (Kukk 2003) are reported. The vegetation composition of localities of L. sibirica populations in the Czech and Slovak Republic is mainly influenced by pH and nutrient gradients of the soil. Those ecological gradients plus the level of water table, amount of calcium carbonate and phosphorus availability are reported by Hájek and Hájková (2011) as the main ecological factors influencing species composition of vegetation in nutrient-poor and rich fens. In particular, rich fens are characterized by high pH values, calcium and low phosphorous availability (Hájek & Hájková 2011), and they also tend to have lower C:N ratios than nutrient-poor fens (Succow 1988, Ellenberg & Leuschner 2010). Indeed, soil pH was the strongest predictor of vegetation composition in our data. Highest pH values and highest amounts of calcium in the soil are in Slovak habitats (SK SR 1-8) which is due to the calcium rich bedrock such as limestone and dolomite (Mello 2000). The only exception is the locality in NP Nízké Tatry (SK N/1) which is on acidic metamorphic substrate (Biely 1992). In the Czech habitats, the geological substrate is sandstone with calcium inclusions (Česká geologická služba 1993, 1995) with the exception of the locality in Pošumaví in South Bohemia (CZ S/1) which is on acidic metamorphic substrate (Česká geologická služba 1991). However, the lowest C:N ratios and highest N contents were not found in the calcium-rich (Slovak) fens, but in deteriorated Czech habitats with the dominance of grasses such as Molinia caerulea and Phragmites australis. These were also characterized by the highest amount of nitrogen/nutrients according to Ellenberg values, while the lowest amount was in nutrient-poor Caricion davallianae communities in Slovakia. Drainage of fens leads to aeration of the surface layer of the peat and thus to its mineralization, thereby releasing nutrients (e.g. Okruszko 1993; Stanová 2000; Holden et al. 2004).

90 Chapter III

Calcium concentration was positively and phosphorus concentration negatively correlated with the poor-rich gradient of fen types according to Succow (1988), Hájek et al. (2006) and Ellenberg and Leuschner (2010) in the aboveground biomass, but not in the soil. This is, because phosphorous forms oxides of changing solubility (Perez-Corona et al. 1996; Blume et al. 2010), and in calcium rich sites its availability is often strongly limited (Wassen et al. 2005; Rozbojová & Hájek 2008). A similar pattern we assume for calcium, which is mainly present as solid, not dissolved particles of calcium carbonate in varying amounts in the soils. Contrary to our expectation soil nutrients were a better predictor to characterize vegetation composition of the L. sibirica habitats than nutrients in leaf biomass. Possible reasons may be firstly an overwhelming effect of pH. Second, there is a strong seasonal variability of N and P contents in biomass (e.g. Jaeger et al. 1999, Mysterud et al. 2011). Still, this is probably a minor effect because we sampled all the samples within one month. Third, a higher nutrient uptake with better nutrient availability may directly lead to higher biomass production in L. sibirica (see below), or may have been stored belowground (McJannet et al. 1995), and thus may have not been reflected in higher concentrations of nutrients in the leaves. Finally, the whole aboveground biomass, i.e. from all species present at a site, is usually sampled when nutrient availability of fens is estimated (Güsewell & Koerselman 2002). This may lead to more clear results, because nutrient concentrations vary widely among plant species inhabiting habitats of different nutrient status (McJannet et al. 1995, Aerts & Chapin 2000). However, a bias is also possible in that approach due to different plant species which may grow in habitats with similar nutrient status. Moreover, the purpose of our study was to assess how the different nutrient status of localities influences individuals of L. sibirica.

Effects of environmental variables on morphological characteristics of the plants As we described in the previous chapter, Ligularia sibirica is a fen species with the ability to grow and/or persist in a variety of habitats. This may be possible due to the high morphological variability of the species. We observed a large morphological variability, both within and among populations. Within population variability is manifested mainly in the shape of the leaf blade, pubescence, and colour of leaf and flowering stem (Krasnoborov1997; Hendrych 2003). Although, such variability of L. sibirica led in the main distribution area, in central Russia, to the description of several species and subspecies (Krasnoborov1997; Hendrych 2003), this kind of classification is not accepted in the European outposts (Hendrych 2003; Kobiv 2005). Among population variability is manifested mainly in the size of individuals. This feature was in past repeatedly reported as a difference between sizes of individuals which grow in full light and moderate shade of shrubs and trees. Specifically, in moderate shade individuals are higher than on full light (Hendrych 2003; Kukk 2003).

91 Chapter III

In our study we tested the effect of selected environmental variables on the morphology of L. sibirica. We found that the basic morphological measures of the species are mainly influenced by habitat quality, Ellenberg indicator values for nutrient availability and light, C:N ratio in the soil and soil pH. When looking at the leaf biomass, N and Ca concentrations were good predictors for some morphological measures, but similarly – and for the same reasons – as described for the vegetation composition, they did not generally explain individual performance of L. sibirica better than soil nutrients. Individuals of L. sibirica are smaller (e.g. low number of rosettes and flowers, seed production, length of flowering stems) under full light, low nutrient availability according to Ellenberg, higher C:N ratio in the soil and higher soil pH. Under these conditions, they accumulate higher amounts of calcium in their leaf biomass, while N concentrations in the leaves are lower. Such features correspond with the features of low productive Caricion davallianae communities (Grootjans et al. 2005). Fen plant species which grow in this habitat are adapted to nutrient-poor conditions (Ventterink et al. 2003) and low competition (Lepš 1999). Although the individuals of L. sibirica at those sites are smaller and have lower seed production, the populations of the species are not endangered. The opposite is true: The presence of all life stages of the species is more important for the viability of whole population than the size of individuals (Heinken-Šmídová & Münzbergová, in press). Individuals of L. sibirica are bigger (e.g. high number of rosettes and flowers, seed production and length of flowering stems) in the shadow of surrounding vegetation, higher nitrogen/nutrient availability according to Ellenberg, low C:N ratio in the soil and lower pH and Ca content in the soil. Under these conditions, they accumulate higher amounts of N in their leaf biomass, while calcium concentrations in the leaves are lower. Such features correspond to former fens which had been meliorated in the past or to habitats after cessation of traditional use. Release of nutrients following drainage (e.g. Okruszko 1993, Stanová 2000, Holden et al. 2004) increases biomass production (Rydin & Jeglum 2006). Mineralization of peat leads to the release of CO2, while N is mainly incorporated into the microbial biomass (e.g. Blume et al. 2010). Consequently, the C:N ratio decreases with humification of the peat (Malmer & Holm 1984). Nutrient availability and productivity of peatlands generally increase with decreasing C:N ratio (Rydin & Jeglum 2006; Succow 1988, Leuschner & Ellenberg 2010). High biomass production at these sites is directly related to a higher uptake of nitrogen and phosphorus by the vegetation. Thus L. sibirica, unlike many other species of nutrient-poor habitats (Pegtel 1994), has the ability to use a higher nutrient supply to grow and compete successfully under various nutrient regimes. Higher size of individuals in the productive habitats can be, however, explained not only by higher stature of the individuals as a response to nutrient availability but also by the fact that all individuals in these localities are old and thus large (Heinken-Šmídová & Münzbergová, in press).

92 Chapter III

Smilar effects on fen ecosystems as drainage may have abandonment of habitats. Though abandoment of habitats leads to the accumulation of biomass and litter, it does not necessarily cause increase of biomass production (Diemer et al. 2001). Both drainage and abandonment of fens lead to a decline of species richness and a change in dominance of species, especially from sedges to grasses (Diemer et al. 2001; Pauli et al. 2002; Mälson et al. 2008). Concretely, the result is often an expansion of grass species such as Molinia caerulea which has the ability to effectively use phosphorus and is a strong competitor (Mälson et al. 2008; Hájková et al. 2009). Under such conditions M. caerulea creates tussocks which strongly compete for light with small rosettes forbs (Diemer et al. 2001). In such habitats adult individuals of L. sibirica are able to compete with tussock forming perennial grasses for nutrients and light. In contrast to adult plants, seedlings and juveniles of L. sibirica are competitively weak and are thus missing in localities with dominant M. caerulea (Heinken-Šmídová & Münzbergová, in press), probably due to the strong competition for light and water. Such response of seedlings and juveniles was shown in experiment for several species by Kottorová & Lepš (1999). A similar contrasting effect of increased productivity, i.e. lower seedling survival and larger adult individuals, is reported by Colling et al. (2002) for Scorzonera humilis. Consequently, most populations which persist in those degraded habitats are of small size.

Conclusions Our results together with those from Heinken-Šmídová & Münzbergová (in press) indicate that although Ligularia sibirica has the ability to grow or survive in a variety of wetland habitats, viable populations of this tall forb are at least in the Czech and Slovak Republic mainly in Caricion davallianae communities or in the transition between Caricion davallianae and Molinion caeruleae communities. Though the individuals of L. sibirica are in those communities of small sizes with low number of leaf rosettes, flower heads and seeds, there are suitable ecological conditions for all life stages of the species such as a sufficient water level, sufficient light conditions and low competition. In contrast, most populations of L. sibirica which persist in degraded or abandoned habitats are on first view impressive by the size of flowering individuals. Though they are apparently able to compete with tall growing herbs and grasses, the viability of those populations is low because of missing early life stages of the species such as seedlings and juvenile individuals and consequently also the population size is small. We conclude that for the viability of L. sibirica the presence of all life stages is more important than the size and robustness of single individuals. Furthermore, the seeming effect of increased productivity on growth of adult individuals underlines the importance of a detailed demographic study of L. sibirica to assess population viability and to give advices for long-term protection of the species.

93 Chapter III

Acknowledgements We would like to thank to the workers in the protected areas of the Czech and Slovak Republic where L. sibirica occurs for providing us with information about its localities. We thank to Thilo Heinken for his help with determination of the plant communities of Ligularia sibirica. We obtained permission from the Ministries of the Environment of the Czech and Slovak Republic to enter the localities of populations of this species, and for its manipulation. This study was supported by grant MŠMT 2B06178 and also partly by grants MŠMT 0021620828 and AV0Z60050516.

References Aerts, R. & Chapin, F.S. III (2000): The mineral nutrition of wild plants revisited: a re-evaluation of processes and patterns. Adv. Ecol. Res., 30: 1–67. Biely, A. (ed.) (1992): Geologická mapa Nízkych Tatier. In: Regionálne geologické mapy Slovenska, 1: 50 000. – Štátny geologicky ústav Dionýza Štúra, Bratislava. Bergamini, A. & Pauli, D. (2001): Effects of increased nutrient supply on bryophytes in montane calcareous fens. J. Bryol., 23: 331-339. Blume, H.P., Brümmer, G.W., Horn, R., Kandeler, E., Kögel-Knabne,r I., Kretzschmar, R., Stahr, K. & Wilke, B.M. (2010): Scheffer/Schachtschabel: Lehrbuch der Bodenkunde. 16. ed. Spektrum Akademischer Verlag, 540 pp. Boeye, D., Verhangen, B., van Haesebroeck, V. & Verheyen, R.F. (1997): Nutrient limitation in species-rich lowland fens. J. Veg. Sci., 8: 415-424. Broz, E. & Przemyski, A. (1988): Stanowisko Ligularia sibirica (L.) Cass. na obszarze Niecki Wloszczowskiej. (The locality of Ligularia sibirica (L.) Cass. in the Wloszczowska Trough (Central Poland).) Fragm. Flor. Geobot., 31-32: 295-299. Brys, R., Jaquemyn, H., Endels, P., van Rossum, F., Hermy, M., Triest, L., de Bruyn, L. & de Blust, G. (2004): Reduced reproductive success in small populations of the self-incompatible Primula vulgaris. J. Ecol., 92: 5-14. Colling, G., Matthies, D. & Reckinger, C. (2002): Population structure and establishment of the threatened long-lived perennial Scorzonera humilis in relation to environment. J. Appl. Ecol., 39: 310-320. Council of European Communities (1992): Council Directive 92/43/EEC of 21 May on the conservation of natural habitats and of wild fauna and flora. Off. J. Eur. Communities, 35: 7– 50. Česká geologická služba (1991): Geologická mapa ČR, Český Krumlov. List 32-23. [1:50 000. edice Geologická mapa ČR 1:50 000]. Česká geologická služba (1993): Geologická mapa ČR, Česká Lípa. List 02-42. [1:50 000. edice Geologická mapa ČR 1:50 000]. Česká geologická služba (1995): Geologická mapa ČR, Mladá Boleslav. List 03-33. [1:50 000. edice Geologická mapa ČR 1:50 000]. Diekmann, M. (2003): Species indicator values as an important tool in applied plant ecology – a review. Basic Appl. Ecol., 4: 493–506. Diemer, M., Oetiker, K. & Billeter, R. (2001): Abandonment alters community composition and canopy structure of Swiss calcareous fens. Appl. Veg. Sci., 4: 237-246. Doust, L.L. (1981): Population dynamics and local specialization in a clonal perennial (Ranunculus repens): I. The dynamics of ramets in contrasting habitats. J. Ecol., 69: 743–755. Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W. & Paulißen, D. (1992): Zeigerwerte von Pflanzen in Mitteleuropa. 2nd ed. Scripta Geobotanika, 18: 1-258. Ellenberg H. & Leuschner C. (2010): Vegetation Mitteleuropas mit den Alpen. UTB, Ulmer. European Communities (2007): LIFE and Europe’s wetlands: Restoring a vital ecosystem. Office for Official Publications of the European Communities Luxembourg, pp. 68.

94 Chapter III

Feráková, V., Maglocký, Š., Marhold, K. (2001): Červený zoznam papraďorastov a semenných rastlín Slovenska (december 2001). In Baláž, D., Marhold, K., Urban, P. (eds) Červený zoznam rastlín a živočíchov Slovenska. Ochr Prír, Banská Bystrica, 20: 48–81. Grootjans, A., Alserda, A., Bekker, R., Janáková, M., Kemmers, R., Madaras, M., Stanová, V., Ripka, J., Ven Delft, B. & Wolejko, L. (2005): Calcareous spring mires in Slovakia; Jewels in the Crown of the Mire Kingdom. In: Steiner, G.M. (ed.): Mires from Siberia to Tierra del Fuego. Stapfia 85: 97-115. Güsewell, S. & Koersalman, W. (2002): Variation in nitrogen and phosphorus concentrations of wetlands plants. Perspect. Plant Ecol. Evol. Syst., 5: 37-41. Hájek, M., Horsák, M., Hájková, P. & Dítě, D. (2006): Habitat diversity of central European fens in relation to environmental gradients and effort to standardise fen terminology in ecological studies. Perspect. Plant Ecol. Evol. Syst., 8: 97-114. Hájek, M. & Hájková, P. (2011): Vegetace slatin, přechodových rašelinišť a vrchovištních šlenků (Scheuchzerio palustris-Caricetea nigrae). In Chytrý ,M. (ed.) Vegetace České republiky 3. Vodní a mokřadní vegetace / Vegetation of the Czech Republic 3. Aquatic and Wetland Vegetation. Academia, Praha, 614-704. Hájková, P., Hájek, M. & Kintrová, K. (2009): How can we effectively restore species richness and natural composition of a Molinia-invaded fen? J. App. Veg., 46: 417-425. Hartig, E.K., Grozev, O. & Rosenzweig, C. (1997): Climate change, agriculture and wetlands in Eastern Europe: vulnerability, adaptation and policy. Clim. Chang., 36: 107-121. Heinken-Šmídová, A. & Münzbergová, Z. (in press): Population dynamics of the endangered, long-lived perennial species, Ligularia sibirica. Folia Geobot. Hendrych, R. (2003) Poznatky o druhu Ligularia sibirica v Čechách. Preslia, 75: 39–69. Holden, A., Chapman, P.J. & Labadz, J.C. (2004): Artificial drainage of peatlands: hydrological and hydrochemical process and wetland restoration. Prog. Phys. Geogr., 28: 95–123. Jaeger, C.H., Monson, R.K., Fisk, M.C. & Schmidt, S.K. (1999): Seasonal partitioning of nitrogen by plants and soil microorganisms in an alpine ecosystem. Ecol., 80: 1883–1891. Käfer, J. & Witte, M.J.P. (2004): Cover-weighted averaging of indicator values in vegetation analyses. J. Veg. Sci., 15: 647–652. Kobiv, Y. (2005): Ligularia sibirica (L.) Cass. (Asteraceae) in the Chornohora mountains (Ukrainian Carpathians): population-ontogenetic parameters, morphology, taxonomy and conservation. Ukr. Bot. Zhurn., 62: 383-395. Kottorová, I. & Lepš, J. (1999): Comparative ecology of seedling recruitment in an oligotrophic wet meadow. J. Veg. Sci., 10: 175–186. Králová, M. (1990): Vybranné chemické analýzy půd a rostlin. Studie ČSAV 1990/12, Praha. Krasnoborov, I.M. (ed.) (1997): Flora Sibiriae, Tomus 13 Asteraceae (Compositae). Novosibirsk, Nauka 169-174 pp. Kukk, Ü. (2003) The distribution of Ligularia sibirica (L.) Cass. in Estonia and changes in its population. Biul. Ogrodów Bot., 12: 11–22. Lepš, J. (1999): Nutrient status, disturbance and competition: an experimental test of relationships in a wet meadow. J. Veg. Sci., 10: 219-230. McJannet, C.L., Keddy, P.A. & Pick, F.R. (1995): Nitrogen and phosphorus tissue concentrations in 41 wetland plants: A comparison across habitats and functional groups. Funct. Ecol., 9: 231-238. Malmer, N. & Holm, E. (1984): Variation in the C/N-quotient of peat in relation to decomposition rate and age determination with 210Pb. Oikos, 43: 171-182. Mälson, K., Backéus, I. & Rydin, H. (2008): Long term effects of drainage and initial effects of hydrological restoration on rich fen vegetation. Appl. Veg. Sci., 11: 99-106. Mattauch,F. (1936): Über die Ligularia sibirica (L.) Cass.. Natur Heimat, Aussig 7: 57–62. Mello, J. (ed.) (2000): Geologická mapa Slovenského raja, Galmusu a Hornádskej kotliny. In: Regionálne geologické mapy Slovenska, 1: 50 000. – Štátny geologicky ústav Dionýza Štúra, Bratislava. Meusel, H. & Jäger, J. (1992a, 1992b): Vergleichende Chorologie der zentraleuropäischen Flora. Vol. 3. (1992a) Text, (1992b) Karten. G. Fischer, Jena. 491p. Moravec, J. (ed.) (1994): Fytocenologie, Academia, Praha, 403 p.

95 Chapter III

Mysterud, A., Hessen, D.O., Mobæk, R., Martinsen, V., Mulder, J., Austrheim, G. (2011): Plant quality, seasonality and sheep grazing in an alpine ecosystem. Basic Appl. Ecol., 12: 195- 206. Okruszko, H. (1993): Transformation of fen-peat soils under the impact of draining. Zeszyty Problemowe Postepow Nauk Rolniczych, 406: 3-73. Olaczek, R. (2004): Ligularia sibirica (L.) Cass., Języczka syberyjska. W: Sudnik-Wójcikowska B., Werblan-Jakubiec H. (red.) 2004. Gatunki roślin. Poradniki ochrony siedlisk i gatunków Natura 2000 – podręcznik matodyczny. 9: 140-144. Ministerstwo Środowiska, Warszawa. Operea, A. & Culiţă, S. (2010): Phytocoenotic surveys on some mesotrophic – eutrophic marsches in eastern Romania. J. Plant Develop., 17: 75-108. Pauli, D., Peintinger, M. & Schmid, B. (2002): Nutrient enrichment in calcareous fens: effects on plant species and community structure. Basic Appl. Ecol., 3: 255–266. Pegtel, D.M. (1994): Habitat characteristics and the effect of various nutrient solutions on growth and mineral nutrition of Arnica montana L. grown on natural soil. Plant Ecol., 114: 109-121. Peintinger, M. & Bergamini, A. (2006): Community structure and diversity of bryophytes and vascular plants in abandoned fen meadows. Plant Ecol., 185: 1-17. Perez-Corona, M.E., van der Klundert, J. &. Verhoeven, J.T.A. (1996): Availability of organic and inorganic phosphorus compounds as phosphorus sources for Carex species. New Phytologist, 133: 225–231. Procházka, F. (ed) (2001): Černý a červený seznam cévnatých rostlin České republiky (stav v roce 2000). Příroda, Praha 18: 1–166. Rode, M.W. (1995): Aboveground nutrient cycling and forest development on poor sandy soil. Plant Soil, 168-169: 337-343. Rydin, H. & Jeglum, J.K. (2006): The biology of peatlands. Oxford University press.343 pp. Schaffers, A.P. & Sýkora, K.V. (2000): Reliability of Ellenberg indicator values for moisture, nitrogen and soil reaction: a comparison with field measurements. J. Veg. Sci., 11: 225-244. Shine, C. & de Klemm, C. (1999): Wetlands, water, and the law: using law to advance wetland conservation and wise use. IUCN environmental policy and law paper, 38, Gland, IUCN, pp. 332. Slavík, B. (2004): Ligularia Cass. In Slavík, B., Štěpánková, J. (eds) Květena České republiky, Vol. 7. Academia, Praha, pp 306–309. S-Plus (2000): Professional edition for Windows, release 2. (1999). MathSoft, Inc., Massachusetts Stančić, Z., Žganec, K. & Gottstein, S. (2010): Marshland vegetation of Plitvice Lakes National Park (Croatia). Candollea, 65: 147-166. Stanová, V. (ed) (2000): Rašeliniska Slovenska. Daphne – Inštitút aplikovanej ekológie, Bratislava, 194 p. Šmídová, A., Münzbergová, Z. & Plačková, I. (2011): Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae). Flora, 206: 151–157. Succow, M. (1988): Landschaftsökologische Moorkunde. Jena: G. Fischer, 340 pp. Ter Braak, C.J.F. & Šmilauer, P. (2002) CANOCO Reference manual and CanoDraw for Windows User’s Guide: Software for canonical community ordination, Version 4.5. Microcomputer Power, Ithaca. Van Duren, I.C., Boeye, D. & Grootjans, A.P. (1997): Nutrient limitations in an extant and drained poor fen: implications for restoration. Plant Ecol., 133: 91-100. Van Duren, I.C. & Pegtel, D.M. (2000): Nutrient limitations in wet, drained and rewetted fen meadows: evaluation of methods and results. Plant Soil, 220: 35-47. Venterink, H.O., Wassen, M.J., Verkroost, A.W.M. & de Ruiter, P.C. (2003): Species richness- productivity patterns differ between N-, P-, and K-limited wetlands. Ecol., 84: 2191-2199. Verhoeven, J.T.A., Kooijman, M.M. & Wirdum, G. (1988): Mineralization of N and P along a trophic gradient in a freshwater mire. Biogeochem., 6: 31-43. Wassen, M.J., Venterink, H.O., Lapshina, E.D. & Tanneberger, F. (2005): Endangered plants persist under phosphorus limitation. Nature, 437: 547-550. Zbíral, J. (1995): Analýza půd I, Jednotné pracovní postupy, SKZÚZ Brno.

96 Chapter III

______

0.45 0.45

0.58

Length of Length of Length of No. of flower No. of

0.48 0.68 0.61 0.68 0.81 0.67 0.68 0.67 0.67

. Significant r values are in bold, P < 0.05 P < bold,are values in r . Significant 0.74 0.64 0.94 0.38

Ligularia sibirica Ligularia 0.62 0.87 0.72 0.60 0.72 0.46

0.59

0.09 0.44

morphological characteristics of characteristics morphological 0.86

No. of rosettes individuals of vegetative No. of leaves 0.18 of flowering individuals individuals of flowering No. of rosettes 0.36 of flowering individuals individuals of flowering stems No. of flowering -0.35 -0.28 Length of flowering stems stems Length of flowering 0.13 0.31 Length of inflorescence Length of inflorescence heads No. of flower -0.28 -0.24 -0.33 -0.34 0.41 0.25 per inflorescence per flower No. of seeds per individual No. of seeds -0.36 -0.32 -0.17 -0.21 0.48 -0.02 0.03 0.54 0.48 0.35 No. of No. of No. of No. of No. of individuals individuals individuals individuals leaves of vegetative rosettes of vegetative leaves of flowering rosettes of flowering flowering stems flowering inflorescence per heads stems per seeds inflorescence flower head Electronic Supplementary Material between relationship The Table S1: ______

97 Chapter III

Table S2: The relationship among nutrient contents in the soil. Significant r values are in bold, P < 0.05

C org. soil N total soil Ca soil K soil [mg/kg] P soil C:N soil [%] [%] [mg/kg] [mg/kg]

N total soil [%] 0.51 Ca soil [mg/kg] 0.86 0.42 K soil [mg/kg] 0.09 0.77 -0.10 P soil [mg/kg] 0.49 0.07 0.27 0.04 C:N soil 0.39 -0.52 0.28 -0.61 0.27 pH 0.31 -0.47 0.10 -0.57 0.15 0.83

Table S3: The relationship among nutrient contents in the leaf biomass. Significant r values are in bold, P < 0.05

C total N total Ca biomass K biomass P biomass

biomass [%] biomass [%] [mg/kg] [mg/kg] [mg/kg] N total biomass [%] 0.29 Ca biomass [mg/kg]] 0.17 0.12 K biomass [mg/kg] 0.56 -0.14 -0.26 P biomass [mg/kg] -0.03 -0.02 0.22 0.21 C:N Biomass -0.19 -0.98 -0.11 0.25 0.09

Table S4: The relationship between Ellenberg indicator values and the habitat quality (scores of samples on the first DCA axis). Significant r values are in bold, P < 0.05

Ellenberg L Ellenberg N Ellenberg M Ellenberg R

Ellenberg N -0.77 Ellenberg M 0.30 -0.45 Ellenberg R 0.05 0.34 -0.37 habitat quality -0.86 0.93 -0.30 0.14

98

CHAPTER IV

Habitat conditions are better determinants of population performance of the perennial herb Ligularia sibirica than population size and genetic diversity

Heinken-Šmídová A. & Münzbergová Z. [manuscript]

Chapter IV

Abstract Understanding factors affecting population performance is a key prerequisite for setting efficient management plans. Habitat conditions, population size and genetic diversity of a population are the main factors. Because these factors are usually studied separately, their relative importance is still largely unknown. We performed a set of regression analyses to detect the pure and overall effect of habitat conditions, population size and genetic diversity on single vital rates and on population performance of a wetland long-lived perennial plant species Ligularia sibirica, in the Czech and Slovak Republic. Our findings indicate that the population growth rate of L. sibirica and the importance of single transitions to population growth rate are mainly affected by habitat conditions. In contrast, single vital rates were affected by all three studied factors. In particular, genetic diversity and population size mainly affected growth related transitions, whereas habitat conditions mainly affected germination, stasis and mortality. The results indicate that the effect of a given factor on single vital rate does not necessarily mean that the same factor also has an effect on overall population growth rate. The results also show that in spite of being partly correlated, habitat conditions, population size and genetic diversity may have independent effects and that we need to consider all of these to understand the possible determinants of plant performance.

Keywords: Habitat quality, Long-lived plant species, Population viability, Single vital rate

Introduction An important aim of plant conservation research is to understand which factors affect the population viability of rare and threatened plant species and how (Primack 2002; Heywood & Iriondo 2003). Numerous previous studies emphasized three main factors affecting the viability of populations of rare species: habitat quality, population size and genetic diversity (e.g. Schemske et al. 1994; Oostermeijer et al. 2003; Vergeer et al. 2003). The effect of all of the above mentioned factors on the long-term persistence of plant populations may strongly differ between species as well as between populations of the same species (Fischer & Matthies 1998; Colling et al. 2002; Brys et al. 2005; Schleuning & Matthies 2009). Habitat quality, population size and genetic diversity are often correlated with each other (e.g. Fischer & Matthies 1998; Oostermeijer et al. 2003; Vergeer et al. 2003; Leimu et al. 2006; Ouborg et al. 2006; de Vere et al. 2009), but each of them may also have some independent effects (Ouborg et al. 2006). Changes in habitat quality are often mentioned as being the most important factors influencing population viability and the long-term persistence of plant populations (Brys et al.

100 Chapter IV

2005; Schleuning & Matthies 2009). In wetlands and grassland ecosystems they are often related to changes in nutrient and water availability (Colling et al. 2002; Brys et al. 2005) and/or disturbance regimes (Oostermeijer et al. 1996). Negative changes in habitat quality often lead to the lower viability of populations (Schemske et al. 1994), which consequently leads to a reduction in population size (Ouborg et al. 2006). Changes in habitat quality were previously shown to affect single vital rates such as seed production (Oostermeijer et al. 1998; Brys et al. 2005), germination (Colling et al. 2002, Vergeer et al. 2003; Brys et al. 2005) and survival rate (Brys et al. 2005), as well as population growth rate (Brys et al. 2005; Colling & Matthies 2006; Schleuning & Matthies 2009). Changes in population size are either a consequence of reduced individual fitness or direct destruction of part of the population (Oostermeijer et al. 2003; Leimu et al. 2006) and they are often related to changes in habitat quality. In the case of perennial species, there may be a long time lag between the changes in habitat conditions and the changes in population size (Hanski & Ovaskainen 2002; Colling & Matthies 2006). In turn, reduced population size leads to the loss of genetic diversity due to random drift or inbreeding (Barrett & Kohn 1991; Oostermeijer et al. 2003) and to various Alee effects, such as a disruption in the interaction between plants and pollinators (Ågren 1996). These changes in population size are reported to have similar effects on particular single vital rates, such as seed production (Paschke et al. 2002; Vergeer et al. 2003; Münzbergová 2006a), germination rate (Vergeer et al. 2003), the survival of adults (Luijten et al. 2000) and population growth rate (e.g. Fischer & Matthies 1998; Münzbergová 2006a), as changes in habitat quality. Reduction in genetic diversity is often a consequence of a reduction in population size (see above; reviewed in Oostermeijer et al. 2003; Leimu et al. 2006). In the short term, a reduction in genetic diversity is associated with lower fitness (e.g. low seed production and germination rates, high seedling mortality and poor growth of seedlings) (Ellstrand & Elam 1993; Oostermeijer et al. 2003) and is consequently expected to result in lower population growth rates and higher extinction probabilities in small populations (Fischer & Matthies 1998). In the long term, it may also reduce the ability of a species to adapt to changing environmental conditions and thus lead to extinction of the whole population (Young et al. 1996; Fischer & Matthies 1998; Ouborg et al. 2006). While all the three factors (i.e. habitat quality, population size and genetic diversity) have been extensively studied in the past, very few studies considered all of them together (de Vere et al. 2009). As a result, it is not possible to identify, which of these three factors is the most important in each specific case. Furthermore, most of the existing studies looked at the correlations between these factors and either single vital rates of plants, such as seed production, germination or flowering probability (e.g. Luijten et al. 2000; Paschke et al. 2002; Oostermeijer et al. 2003; de Vere et al. 2009), or population dynamics (Dahlgren & Ehrlén 2009). However, it was shown that the effects of the factors studied on single vital rates could be combined (Vergeer et al. 2003; de

101 Chapter IV

Vere et al. 2009) and that differences in single vital rates may not imply differences in the whole life cycle and vice versa (Münzbergová 2006a; Kolb et al. 2010). Studies that investigated the possible effects of habitat quality, population size and genetic diversity on single vital rates and population performance in one model are, however, largely missing (but see Vergeer et al. 2003; de Vere et al. 2009). These studies only looked at the effects on single vital rates and not on overall population dynamics. Both Vergeer et al. (2003) and de Vere et al. (2009) used path analyses thus including possible interactions between the studied factors. They found that the reduced fitness of Succisa pratensis and Cirsium dissectum is better explained by genetic effects and habitat deterioration, respectively, than by population size. Here, we investigate how habitat conditions, population size and genetic diversity influence the population performance of Ligularia sibirica (L.) Cass., a long-lived, perennial wetland herb. Ligularia sibirica is a postglacial relict wetland plant species strictly protected by the Habitats Directive, Annex II of the Council of the European Community (1992). We followed 11 populations of our study species in the Czech and Slovak Republic over 4 years. The studied populations vary largely in size and grow in a range of different habitats (Hendrych 2003; Šmídová 2006, 2009). In previous studies, we investigated the genetic variability and detailed demography of L. sibirica populations. While the level of genetic diversity in some of the populations was high, this was not true in others (Šmídová et al. 2011). Most of the populations studied are performing well and only those growing in nitrogen-rich habitats have a high extinction risk (Heinken- Šmídová A. & Münzbergová Z., unpublished data). By combining data on the full life cycle with data on habitat conditions, population size and genetic diversity, we asked: what is the effect of year, population size, habitat conditions and genetic diversity on single vital rates and on the whole life cycle of the species? Because the effects of habitat conditions, population size and genetic diversity can be partly correlated, we tested both the overall effects (i.e. the effect of each of the three factors separately) and the pure effects of each of the factors using the other two as covariates.

Material and methods

Study species Ligularia sibirica (L.) Cass. (Asteraceae) is a perennial hemicryptophyte 124.25 ± 43.00 cm high (mean ± SD) with ground rosettes of leaves and a short rhizome. During flowering, from mid- July to the end of August, L. sibirica creates 1.91 ± 1.73 flowering stems (mean ± SD) with 18.79 ± 8.42 flower heads per inflorescence (mean ± SD). The flower heads contain 32.13 ± 7.26 achenes (mean ± SD) (Heinken-Šmídová A. & Münzbergová Z., unpublished data). The achenes (hereafter referred to as seeds) are oval shaped and weigh 2.16 mg ± 0.797 (mean ± SD) (Šmídová et al. 2011). Their pappus forms a tuft of hair longer than the achenes (Hegi 1929; Slavík 2004). The

102 Chapter IV seeds are dispersed over a short distance by either wind or gravity. The terminal velocity of L. sibirica seeds varies between 0.86 and 1.18 m s -1 (Šmídová et al. 2011). There is a pre-dispersal seed predator from the family Gelechiidae (Lepidoptera), with larvae that develop inside the flower head and internally consume the seeds of L. sibirica (Elsner & Patočka, pers. comm.). The species is diploid (2n = 60) (Liu 2004). Ligularia sibirica has hermaphrodite entomogamous flower heads (Slavík 2004). Sexual reproduction is highly predominant. Clonal growth of the ‘phalanx’ type, i.e. in the form of a tightly-packed advancing front of ramets with a slow radial spread (sensu Doust 1981), may also occur. Ligularia sibirica is a plant species that prefers full sunlight, but it also grows under open tree canopies where seed production is lower (Kukk 2003). It occurs in a variety of wetlands, such as humid grasslands, calcareous fens, transitional mires and alluvial woodlands with the groundwater table at the surface (Slavík 2004). Ligularia sibirica is regarded as a postglacial relict (Hendrych 2003; Šmídová et al. 2011), and is classified as a ‘critically endangered’ species in the Czech Republic (Procházka 2001) and as a ‘vulnerable’ species in the Slovak Republic (Feráková et al. 2001).

Study sites Demographic data were collected from nine populations of L. sibirica in the Czech Republic and two populations in the Slovak Republic (Table 1). The study populations in the Czech Republic are located in three regions. The first one (containing five local populations) is in northern Bohemia near the village Jestřebí (50°36’8” - 50°36’22” N, 14°36’51” - 14°37’21” E). The second one (containing three local populations) is in central Bohemia near the village Rečkov (50°29’38” - 50°30’3” N, 14°51’56” - 14°54’40” E). The last population is in southern Bohemia in Pošumaví, by the fishpond ‘Olšina’ (48°47’34” N, 14°6’7” E). In the Slovak Republic, one population was located in NP Slovenský raj. It is an extensive population along the river Hnilec (48°52’32” N, 20°15’46” E). The second population is in Salvatorské lúky, approximately 25 km west of Prešov (49°2’45” N, 20°56’30” E).

Habitat conditions Most of the populations occur in base-rich fens of the alliance Caricion davallianae. However, there are some exceptions. The first exception is one locality in an alluvial floodplain in central Bohemia (CZ C/2), which can be classified as nitrogen-rich Carex beds of the alliance Magnocaricion elatae. The second exception is the locality in southern Bohemia (CZ S/1), which shows a very close similarity to the nitrogen-poor fens, but is calcium-poor and acidic. We classified this as a transition mire of the alliances Caricion fuscae, Molinion caeruleae and Calthion palustris. The last exceptions are localities in northern Bohemia (CZ N/3, CZ N/4 and CZ N/5). Here, the fens had been partly meliorated in the past (Hendrych 2003). Today, these former fens can be classified as nitrogen-rich degradated Molinion (Šmídová 2006).

103 Chapter IV

Table 1: Characteristics of the studied localities of the Ligularia sibirica populations. The populations were divided according to states (CZ and SK) and regions (N, C, S, SR and P). * Values of habitat quality represent scores of the first DCA axis. ** Values of observed heterozygosity (HO) were taken from Šmídová et al. 2011

Estimated pop. size Observed Population Abbrev. Habitat conditions* of adults in 2006 heterozygosity**

Czech Republic (CZ) northern Bohemia (N) Vojenské louky CZ N/1 126 0.61 0.29 Baronský rybník CZ N/2 243 0.74 0.26 Louky u Robeče CZ N/3 193 2.42 0.36 Rákosina u Robeče CZ N/4 323 2.35 0.32 Sluneční dvůr CZ N/5 123 3.49 0.27 central Bohemia (C) Rečkov CZ C/1 ~ 10 000 1.19 0.37 Louky u NV CZ C/2 ~ 10 000 3.12 0.37 Klokočka CZ C/3 ~ 7 000 0.34 0.28 southern Bohemia (S) Olšina CZ S/1 224 0.64 0.25 Slovak Republic (SK) NP Slovenský raj (SR) SK SR ~ 50 000 1.12 0.39 Prešov (P) Salvatorské lúky SK P/1 ~ 2000 0.11 0.31

In order to describe the habitat conditions at each locality, we recorded one to three phytosociological relevés in 5 × 5 m plots in each locality in such a way that the heterogeneity of the locality was covered. Species abundances were estimated by the Braun-Blanquet cover scale (Moravec 1994). An Ellenberg indicator value was assigned for light, nitrogen availability, moisture and soil reaction (Ellenberg et al. 1992) to each plant species in each relevé if the value was available. Both vascular plant species and bryophytes were considered. The differences between relevés within a single locality were quite low, thus we used one representative phytosociological relevé at each locality and for those we thereafter calculated unweighted arithmetic means of Ellenberg indicator values. It should be noted that the Ellenberg N values indicate general nutrient availability rather than nitrogen availability (Schaffers & Sýkora 2000; Diekmann 2003; Ellenberg & Leuschner 2010). Therefore, in the text we will use Ellenberg N as ‘soil nutrient availability’ and increasing Ellenberg N values will be referred to ‘eutrophication’. We used the Ellenberg indicator values to describe habitat conditions because they are widely and successfully used in this respect (Diekmann 2003; Käfer & Witte 2004). They were mainly developed on the basis of field experience and they reflect the ecological behaviour of species. Therefore, the average indicator values do not refer to habitat conditions at a specific moment, but present integration over time (Ellenberg et al. 1992; Schaffers & Sýkora 2000).

104 Chapter IV

Based on the phytosociological relevés, detrended correspondence analyses (DCA ordination analyses) were performed in Canoco 4.5 (Ter Braak & Šmilauer 2002). We used percentage species data with logarithmic transformation and rare species were down-weighted. The scores of vegetation relevés at the first DCA axis were used to express the habitat conditions of the localities in all subsequent analyses. In order to aid interpretation of the gradient of habitat conditions, we tested the correlation between mean Ellenberg indicator values of light, moisture, soil reaction, nutrient availability and the scores of the first DCA axis to see which of the Ellenberg indicator values explained most of the variability in vegetation composition of the sites.

Population size Population size was estimated in 2006. An adult individual of L. sibirica was defined as an individual consisting of 1 to 15 rosettes in a dense cluster that was connected by up to 10 cm long underground rhizomes most likely representing a single genet (Heinken-Šmídová A. & Münzbergová Z., unpublished data). In the case of small populations (< 500 adult individuals), the population size was the sum of all adult individuals in the locality. In the case of large populations (> 500 adult individuals), the population size was estimated as an area of the population multiplied by the average number of adult individuals in ten 1 m2 plots. The plots were located throughout the whole locality in order to cover the different densities of adult individuals in the proportions in which they occurred at the localities (Table 1). Since higher errors can occur in the estimation of large populations, we used logarithmically transformed the population size values.

Genetic diversity

We used the data on observed (HO) and expected (HE) allozyme heterozygosity measured in a genetic study of the same populations (Šmídová et al. 2011) to separate the effects of habitat conditions from those resulting from genetic processes, as suggested by Oostermeijer et al. (2003). Heterozygosity was determined by assaying 20 randomly selected individuals per population for the following five enzyme systems (which produced nine polymorphic loci) AAT (E.C. 2.6.1.1), DIA (E.C. 1.6.99.3), LAP (E.C. 3.4.11.1), SOD (E.C. 1.15.1.1), 6-PGDH (E.C. 1.1.1.44). For the exact electrophoresis procedure and details of the analyses see Šmídová et al. (2011). Of the populations in NP Slovenský raj in Slovakia, only one population was studied for the detailed demography. This had been considered as three independent subpopulations in the genetic study (SK SR/1, SK SR/4 and SK SR/7 in Šmídová et al. 2011). Therefore, we estimated the values of observed (HO) and expected (HE) heterozygosity based on combined data from all three subpopulations.

105 Chapter IV

Demographic data collection and construction of transition matrices The detailed demography observation was conducted over 4 years from 2005 to 2008. In each study population we marked and followed 50 individuals in each stage, where possible. The life cycle of L. sibirica was divided into seven stages: (1) seed bank first year, (2) seed bank second year, (3) seedlings, (4) juveniles, (5) vegetative adults (6) fertile adults and (7) dormant adults (see Fig. S1 in Electronic Supplementary Material (hereafter ESM)). For details on construction of the transition matrices see Heinken-Šmídová A. & Münzbergová Z. (in press).

Analyses of population dynamics Demographic data were examined using transition matrix models (Caswell 2001). In order to describe the performance of each population, we calculated the population growth rate (Caswell 2001) and elasticity (contribution of each matrix element to population growth rate, de Kroon et al. 2000) for each population and transition interval separately. We calculated 95% confidence intervals to estimate the reliability of the population growth rate estimates (see Fig. S2 in ESM) and elasticity (Alvarez-Buylla & Slatkin 1994) by bootstrapping the original data used to derive the transition matrices 10,000 times, as suggested by Efron and Tibshirani (1994) and Caswell (2001), using a MATLAB script developed by Münzbergová (2006a). In order to summarize the information on population growth rate over populations and years, we used the stochastic simulation approach suggested by Caswell (2001) and Sletvold and Rydgren (2007). We drew a sequence of matrices for each set of matrices. Each matrix from the set was drawn at random and with equal probability and we simulated population growth using this matrix sequence. Each simulation was performed for 10,000 one-year intervals. The simulations were performed using a MATLAB script developed in a previous study (Münzbergová 2005). We ran the same stochastic simulations for the bootstrapped matrices described above and thus were able to construct a 95% confidence interval for the stochastic population growth rate. A life-table response experiment (LTRE) with a fixed factorial design was conducted to examine the effects of locality and year on population growth rate, as suggested by Caswell (2001). We calculated the mean matrix (i.e. across all populations and years) and its growth rate. Differences in the growth rates of the matrices of each type from the growth rate of the mean matrix were calculated (see Münzbergová 2007). The LTRE analysis indicates the contribution of each life-cycle transition to the differences between different levels of each factor (population, year). Important life-cycle transitions are those with large positive contributions to some factor levels and large negative contributions to others. Similar to ANOVA, the mean of the treatment effect is zero (Caswell 2001). The significance of the LTRE was estimated using the permutation test as described in Münzbergová (2007), with 10,000 permutations. In order to summarize the information on the elasticity of single matrix elements and on the contribution of single matrix elements to variations in population growth rate, as identified by the

106 Chapter IV

LTRE analysis, we divided the single elements into growth (G), stasis (L) and fecundity (F), following Silvertown and Franco (1993), and summed the elasticity values and LTRE contributions within these three categories. In order to calculate the life span, we used the method suggested by Cochran and Ellner (1992), which derives age-based measures from stage-structured models. The resulting conditional life span is defined as the average age of death for individuals that survive to a certain stage (Cochran & Ellner 1992). In order to represent the population, we decided to use the life span of the mature stage, which has the highest longevity values, i.e. fertile individuals (see Ehrlén & Lehtilä 2002). The method of Cochran and Ellner (1992) only works with a single matrix. In order to obtain conditional total life span data over all of the years studied in each locality we computed the harmonic mean of the life spans for all matrices representing different years within one population as a measure of the average life span of the species in the locality. We then calculated the arithmetic mean of all population estimates to obtain one value for the species studied (see Ehrlén & Lehtilä 2002). All analyses were performed using MATLAB version 7.3.0.267 (Mathworks Inc. 2006).

Statistical analysis The effects of habitat conditions, population size and genetic diversity in terms of observed and expected heterozygosity on the performance of L. sibirica were investigated by a set of regression analyses. Correlation analyses were performed between independent variables first, and then between dependent variables. In the subsequent analyses, the overall effects of each of the three factors (habitat conditions, population size and genetic diversity) were estimated without including the other two factors in the model. Afterwards, we estimated the pure effect of each of the factors by using the other two as covariates. All regression analyses were calculated with both observed and expected heterozygosity as measures of genetic diversity. As both of these results were very similar, for reasons of simplicity we only present the results for observed heterozygosity. We chose observed heterozygosity because most of the demographic and population viability studies that incorporated genetic information in the analyses used values of observed heterozygosity (e.g. Oostermeijer et al. 2003) and because in our previous study we found that other genetic measurements (i.e. the mean number of alleles per locus, the percentage of polymorphic loci, the inbreeding coefficient and expected heterozygosity) are correlated with observed heterozygosity (Šmídová et al. 2011). Statistical analyses were carried out in S-Plus (2000).

107 Chapter IV

Single vital rates First, we tested the effect of year, habitat conditions, population size, genetic diversity and the interactions between habitat conditions, population size and genetic diversity with year on the number of initiated and developed seeds (including those subsequently damaged by herbivores) and on the proportion of seeds damaged by herbivores using generalized linear models with Gaussian distribution. The numbers of initiated and developed seeds were normally distributed. The proportion of damaged seeds was squared root transformed to gain normality. In the same way, we tested the effect of habitat conditions, population size and genetic diversity on the single vital rates directly included in the transition matrices: the probability of germination, stasis (i.e. where an individual stays at the same stage), growth (i.e. transition to a larger stage), flowering and mortality. The tests were performed separately for seedlings, juveniles, vegetative and fertile adults. Dormant individuals were excluded from the analyses due to their low number. Overall and pure effects of habitat conditions, population size and genetic diversity on single vital rates were tested using logistic regression. The effect of year was only rarely significant. Also, there was only a single marginally significant interaction between year and a factor. The effect of year and the interactions with year are always included in models but not reported.

Population performance and expected life span We tested the effect of habitat conditions, population size and genetic diversity on the stochastic population growth rate estimated by combining all three transition matrices for each population. Furthermore, we tested the effect of these factors on the elasticity of growth (G), stasis (L) and fecundity (F) and on the results of the life-table response experiment (LTRE) summarizing the contributions of growth (G), stasis (L) and fecundity (F) for each population. We also tested the effects of the factors on the expected life span of the individuals in each population. The life span was log transformed. The overall and pure effects of habitat conditions, population size and genetic diversity were tested using generalized linear models with Gaussian distribution.

Results

Habitat conditions The first DCA axis (Fig. S3 in ESM) used to describe habitat conditions of the sites in the subsequent analyses represented a gradient from tall and mostly expansive species of nutrient-rich habitats such as Calamagrostis epigeios, Carex acutiformis, Cirsium oleraceum, Eupatorium cannabinum and Phragmites australis to calcareous fens species such as Carex davalliana, Parnassia palustris, Crepis paludosa and Succisa pratensis. The differences in vegetation composition between the localities were significantly explained by Ellenberg indicator values for

108 Chapter IV nutrient availability (r = -0.94; P < 0.001) and light (r = 0.86; P < 0.001) (Fig. S4 in ESM). The nutrient availability and light variables were strongly correlated (see Table S1 in ESM). In the following text, we describe the habitat condition gradient as a gradient from low quality habitats (nutrient-rich localities dominated by tall species) to high quality habitats (nutrient-poor habitats dominated by calcareous fen species).

Correlation between variables There was a significant correlation between two of the independent variables: population size and genetic diversity (r = 0.65; P < 0.05) (Table 2). There were also many significant correlations between the dependent variables (Table S2 in ESM). Hence, it is not always possible to fully separate one from the other (e.g. the mortality of seedlings and the mortality of fertile adults).

Table 2: Relationship between independent variables i.e. habitat conditions, population size and genetic diversity (Observed heterozygosity, HO). Significance values are in bold, P < 0.05

Habitat Population

conditions size Population size 0.29

HO 0.15 0.65

Single vital rates When looking at the overall effects, the number of initiated and developed seeds and the proportion of damaged seeds were significantly influenced by habitat conditions (Table 3). Specifically, the greatest number of seeds was initiated and then developed in the low quality habitats and the greatest proportion of seeds was damaged by herbivory in the high quality habitat. Additionally, the proportion of damaged seeds significantly increased with population size and with increasing genetic diversity. The effect of habitat conditions on the number of initiated and developed seeds became non-significant when estimating the pure effect. The only remaining significant effect was increasing seed damage with increasing habitat quality. Both overall and pure effects indicate that the germination and stasis of juveniles significantly increased with habitat quality (Table 4). However, the other results differed between the overall and pure effects. When looking at the overall effects, the stasis and mortality of vegetative adults, growth and seedling mortality were significantly influenced by all factors. The probability of stasis for the vegetative adults and the rate of growth of seedlings increased with increasing habitat quality, population size and genetic diversity. The mortality rate of seedlings and vegetative adults decreased with increasing habitat quality, population size and genetic diversity. The mortality rate of juveniles and fertile adults significantly decreased with increasing habitat

109 Chapter IV quality. Also, the mortality rate of fertile adults significantly decreased with increasing genetic diversity. The growth rate of juveniles significantly decreased with increasing population size and with increasing genetic diversity. When looking at the pure effect of the same regression analysis as above, the probability of stasis of juveniles only significantly decreased with increasing genetic diversity (Table 4). The probability of stasis and the flowering of fertile adults increased with both decreasing population size and increasing genetic diversity, while the growth rate of seedlings significantly increased with increasing population size and decreased with increasing genetic diversity. The effect of genetic diversity on the growth rate of seedlings was the only effect that changed direction from positive in the overall effect to negative in the pure effect. The mortality rate of seedlings significantly decreased with increasing population size. The mortality rates of juveniles, vegetative and fertile adults significantly increased with decreasing habitat quality. The mortality rate of juveniles increased with increased genetic diversity, whereas the mortality rate of fertile adults decreased with increased genetic diversity. Growth and flowering of vegetative adults was not significantly influenced by any of the factors studied.

Table 3: Overall and pure effects of habitat conditions, population size and observed heterozygosity (HO) on the number of initiated and developed seeds and on the proportion of damaged seeds. Direction of the effect is ↑ significantly positive and ↓ significantly negative. Df = 1. Df Error = 32. The values show % of the explained deviance. Significance levels: marginally significant in italics, * P < 0.05, **P < 0.01, *** P < 0.001

Observed heterozygosity Habitat conditions Population size (HO) Overall Pure Overall Pure Overall effect effect effect effect effect Pure effect No. of initiated seeds ↓24.71** — — — — —

No. of developed seeds ↓12.35 — — — — —

Proportion of damaged seeds ↑20.19** ↑9.28* ↑27.94*** — ↑11.85* —

Population performance and expected life span The stochastic population growth rate was independent of population size and genetic diversity but it significantly increased with increasing habitat quality when looking at both overall and pure effects (Table 5, Fig. 1). Elasticity was independent of population size and genetic diversity but it was significantly affected by habitat quality (Table 5). The elasticity of growth (G) and fecundity (F) were higher in high quality habitats. On the other hand, the elasticity of stasis (L) was higher in low quality habitats. The effects on elasticity, however, disappeared when looking at the pure effects.

110 Chapter IV

Table 4: Overall and pure effects of year, habitat conditions, population size and observed heterozygosity (HO) on the probability of germination, the probability of stasis in the same stage, and the probability of growth, flowering and mortality. Direction of the effect is ↑ significantly positive and ↓ significantly negative. Df = 1. Df Error = 32 for the probability of germination, stasis in the same stage – vegetative and fertile adults, the probability of growth – vegetative adults, the probability of flowering – vegetative and fertile adults, the probability of mortality – vegetative and fertile adults. Df Error = 21 for the probability of growth – seedlings, for the probability of mortality – seedlings. Df Error = 23 for the probability of mortality – juveniles. Df Error = 25 for stasis in the same stage – juveniles, probability of growth – juveniles. The values show % of the explained deviance. Significance levels: marginally significant in italics* P < 0.05, **P < 0.01, *** P < 0.001

Observed heterozygosity Habitat conditions Population size (HO) Overall Overall Overall effect Pure effect effect Pure effect effect Pure effect Germination ↑37.62*** ↑29.99*** — — — —

Stasis – juveniles ↑49.52*** ↑39.44*** — — — ↓10.05*

Stasis – vegetative adults ↑17.91** — ↑21.61** — ↑8.41 —

Stasis – fertile adults — — — ↓21.54** — ↑19.63**

Growth – seedlings ↑27.65** — ↑40.74*** ↑11.85* ↑27.28** ↓8.23

Growth – juveniles — — ↓14.33* — ↓13.35* — Growth – vegetative — — — — — — adults Flowering – vegetative — — — — — — adults Flowering – fertile adults — — — ↓18.42* — ↑20.71**

Mortality - seedlings ↓30.08** — ↓40.89*** ↓12.56* ↓26.71** —

Mortality – juveniles ↓36.16** ↓31.27* — — — ↑10.346 Mortality – vegetative ↓42.55*** ↓29.28*** ↓20.98*** — ↓6.25* — adults Mortality – fertile adults ↓14.79** ↓29.24*** — — ↓4.38 ↓17.86**

The contribution of growth (G) and stasis (L) to variations in population growth rate, as identified by the LTRE analysis, was significantly influenced by habitat conditions. Additionally, the contribution of growth (G) was significantly influenced by population size (Table 5). The contribution of growth (G) increased with increasing population size and decreased with increasing habitat quality. All of these effects, except for the effect of habitat quality on stasis, disappeared when looking at the pure effects (Table 5). The expected life span of L. sibirica significantly increased with increasing habitat quality and population size. However, when looking at the pure effects, only the effect of habitat quality was significant (Table 5, Fig. 2).

111 Chapter IV

Table 5: Overall and pure effects of habitat conditions, population size and observed heterozygosity (HO) on stochastic population growth rate, on the elasticity of growth (G), stasis (L) and fecundity (F), on the life-table response experiments (LTREs) of growth (G), stasis (L) and fecundity (F) and on the expected lifespan. Direction of effect is ↑ significantly positive and ↓ significantly negative. Df = 1, Df Error = 10. The values show % of the explained deviance. Significance levels: * P < 0.05, **P < 0.01, *** P < 0.001

Observed Habitat conditions Population size heterozygosity (HO) Overall Pure Overall Pure Overall Pure effect effect effect effect effect effect Stochastic population growth rate ↑45.03* ↑39.75* — — — —

Elasticity of growth (G) ↑42.35* — — — — —

Elasticity of stasis (L) ↓46.67* — — — — —

Elasticity of fecundity (F) ↑53.59* — — — — —

LTRE of growth (G) ↓29.58 — ↓44.00* — — —

LTRE of stasis (L) ↑46.35* ↑48.29* — — — —

LTRE of fecundity (F) — — — — — —

Expected lifespan ↑49.05* ↑41.33* ↑23.99 — — —

Discussion We studied the overall and pure effects of habitat conditions, population size and genetic diversity on single vital rates and population performance of the long-lived perennial species Ligularia sibirica in the Czech and Slovak Republic. The overall effect of the studied factors was the effect of one factor without the effects of the other two included in the analyses. Most of the existing studies that considered these factors tested the overall effects as they usually only concentrated on one factor out of the three. In this way, however, they may have falsely assigned the effects of one (unstudied) factor to the effects of another (studied) factor since the three factors are known to be correlated with each other (de Vere et al. 2009). The pure effects estimated the effect of each factor by using the other two as covariates. In this way, they indicated the contribution of a given factor without possible confounding effects from the other factors. A possible disadvantage of using the pure effects can occur in the case of strong correlations between the studied factors. In such a case, a pure effect of none of the factors would be significant, despite significant overall effects. Taking this into account, we present the results of both the overall and pure effects. Studies that used this combined approach are not very frequent. Examples include the studies by Vergeer et al. (2003) on Succisa pratensis and de Vere et al. (2009) on Cirsium dissectum. Both Vergeer et al. (2003) and de Vere et al. (2009) used path analyses thus including possible interactions between the studied factors. They found that the reduced fitness of species was explained by a combination of two factors: genetic diversity and habitat quality. Our result corresponds with these findings. All three factors tested, habitat conditions, population size and

112 Chapter IV genetic diversity had combined effects on various single vital rates. On the other hand, the population growth rate of L. sibirica and the importance of single transitions to population growth rate were only purely affected by habitat conditions.

Single vital rates Habitat quality was the only factor that affected the number of initiated and developed seeds. The absence of a relationship between the numbers of initiated or developed seeds and population size or genetic diversity is in contradiction with the findings of other studies (e.g. Ouborg & van Treuren 1995; Fischer & Matthies 1998; Luijten et al. 2000; Vergeer et al. 2003). For example Fischer and Matthies (1998) found that Gentianella germanica produces a lower amount of seeds in small populations. Such a feature is often explained by either increased inbreeding (e.g. Barrett & Kohn 1991; Fischer & Matthies 1998) or by a disruption in the interaction between plants and pollinators (Ågren 1996), where in small populations, plants may not be attractive to pollinators, resulting in low visitation frequencies (Jennersten 1988; Fischer & Matthies 1998). Alternatively, a lower seed set in small population may be due to the mating system of species, as was found by Brys et al. (2004) and Luijten et al. (2000) for the self-incompatible Primula vulgaris and Arnica montana, respectively. Flowering plants of L. sibirica create two flowering stems with 19 flower heads on average and are thus quite attractive to pollinators, even in small populations. According to the preliminary data, L. sibirica is self-compatible with enhanced seed production after outcrossing (Heinken-Šmídová A. & Münzbergová Z., unpublished data). This fact, together with the ability of selfing, could explain the absence of relationships between seed production and population size and genetic diversity. In our study, the number of seeds was only determined by habitat quality. The number of initiated and developed seeds strongly increased with decreasing habitat quality. This can be explained by the overall higher stature of these plants in habitats of low quality (nutrient-rich habitats), which have more flower heads and consequently produce more seeds. The number of flower heads of L. sibirica ranges from 17 in nutrient -poor to 23 in nutrient - rich habitats, on average. Similarly, Münzbergová (2006a) observed that the number of flower heads increased with increasing above-ground vegetation biomass at the site of Scorzonera hispanica. In contrast to number of initiated and developed seeds, the proportion of damaged seeds was affected by all three factors studied. The proportion of damaged seeds was higher in the larger, genetically more diverse populations and in high quality habitats. The pure effect was, however, only significant for habitat conditions, indicating that the effects of population size and genetic diversity might be due to the correlation of these factors with habitat quality. All three factors have previously been shown to affect the intensity of herbivory (e.g. Kéry et al. 2001; Östergård & Ehrlén 2005; Münzbergová 2006b; Ågren et al. 2008). For example, Kéry et al. (2001) and

113 Chapter IV

Figure 1: Relationship between the scores of vegetation relevés from the localities on the first DCA axis and the stochastic population growth rate (λ). See Table 1 for abbreviations of the study localities

Figure 2: Relationship between the scores of vegetation relevés from the localities on the first DCA axis and the logarithm of expected lifespan. See Table 1 for abbreviations of study localities

114 Chapter IV

Ågren et al. (2008) found higher seed predation in large populations of Gentiana cruciata and Vincetoxicum hirundinaria, respectively, suggesting a higher probability of the loss of associated seed predators in small populations. Several studies have indicated that predispersal seed predators respond positively to increased concentrations of resources, such as the number of flowers and seed production (Cariveau et al. 2004; Östergård & Ehrlén 2005). However, our results contradict these findings (see above). A possible explanation could be that the high quality habitats represented more suitable environmental conditions for the survival of herbivores. Habitat quality had positive effects on germination and stasis and negative effects on mortality. In contrast, population size and genetic diversity mainly had effects on growth-related transitions. The reduced germination rate of populations in low quality habitats could be caused by an increase in productivity of the habitat and the resulting competition for light with the surrounding vegetation (Colling et al. 2002; Brys et al. 2005), leading to a lack of suitable micro- habitats for germination (Vergeer et al. 2003; Brys et al. 2005). In contrast to germination, plant growth (i.e. the growth of seedlings and juveniles) mainly depends on the quality of the plants (Oostermeijer et al. 2003), which is determined by genetic diversity of the population as well as by population size. This feature can be explained by the genetic consequences of small population sizes, such as inbreeding depression, which may cause an immediate loss of fitness (Charlesworth & Charlesworth 1987; Dudash & Fenster 2000). The effects of habitat quality always had the same direction – the plants in more suitable conditions had better germination and more survived. In contrast, the directions of the effects of population size and genetic diversity were more variable. For example, the mortality rate of juveniles increased, while the probability of stasis in juveniles and seedling growth decreased with increasing heterozygosity of the populations, indicating a negative effect of heterozygosity on these early transitions. In contrast, the mortality of fertile adults decreased with increasing heterozygosity, indicating positive effects. Our findings support the evidence indicating that selection pressure can lead to an excess of homozygotes among early life-cycle stages (Oostermeijer et al. 1995) or favour the survival of heterozygotes among adult individuals (Luiten et al. 2000). Similarly, the higher mortality rate in low quality habitats can be related to reduced germination, to an increase in competition with the surrounding vegetation and to light limitation, which was reported, for example, for Primula veris (Brys et al. 2005). Such an explanation might also be true for the higher mortality rate of vegetative adults, as individuals in this life-cycle stage are only formed by ground rosettes. Thus, we assumed that vegetative adults are not good competitors with the surrounding vegetation when the vegetation is high (especially in Molinia caeruela and Phragmites australis dominated habitats). Under these habitats, the nutrient availability presumably increased due to the drainage or accumulation of nutrient -rich sediments. However, the mortality rate of the fertile adults cannot be explained by competition with the

115 Chapter IV surrounding vegetation as the fertile individuals created flowering stems that were over 120 cm high, and thus most individuals were as high as or even higher than the surrounding vegetation. A possible explanation in this case is that the flowering plants were not able to survive in these localities due to other unsuitable conditions, such as the lack of an underground water supply.

Population performance and expected lifespan Population growth rate, individual expected life spans and the contribution of single transitions to population growth rate in L. sibirica were strongly influenced by habitat quality, whereas population size and genetic diversity did not play any role when looking at the pure effect of these factors. However, habitat quality had mixed effects on population size in the case of expected life span and the LTRE of growth. With decreasing habitat quality, the elasticity of stasis increased and the elasticity of growth and fecundity decreased. There was also a strong increase in the population growth rate with increasing habitat quality. Previous published studies found a relationship between population size and habitat conditions and population growth rate. For example, a positive relationship between population size and population growth rate was reported for Gentianella germanica (Fischer & Matthies 1998) and for Scorzonera hispanica (Münzbergová 2006a). Our results are in agreement with the findings of several other studies (e.g. Brys et al. 2005; Colling & Matthies 2006; Schleuning & Matthies 2009) where deterioration in habitat quality led to reduced population growth rates. Our findings thus support the conclusion that nutrient enrichment negatively affects the population performance of perennial herbs (Brys et al. 2005; Colling & Matthies 2006). In our previous study (Heinken-Šmídová A. & Münzbergová Z., unpublished data), we showed that the transition that contributed most to population growth rate was the stasis of individuals. Indeed, this transition was also found to be strongly influenced by habitat quality. This, together with the fact that the conditional life span of fertile adult individuals is lower in populations growing in low quality habitats (mean life span of 77.6 years for populations growing in nutrient -poor habitats and a mean life span of 30.9 years for populations growing in nutrient - rich habitats, Heinken-Šmídová A. & Münzbergová Z., in press), indicates that adult plants may not be able to buffer the negative effects of habitat degradation on populations (Brys et al. 2005). The time delay between a change in habitat quality and population extinction may thus be shortened considerably (Hanski & Ovaskainen 2002; Brys et al. 2005; Schleuning & Matthies 2009). The strong discrepancy between the factors affecting single vital rates and the factors affecting population growth rate is in agreement with the findings of several recent studies (Münzbergová 2006a; Kolb et al. 2010). It supports the idea that identifying the effects of a given factor on single vital rates does not necessarily mean that the same factor has a significant effect on the overall population growth rate.

116 Chapter IV

Conclusions Our results indicate that the population growth rate of L. sibirica and the importance of single transitions for population growth rate are mainly affected by habitat conditions. In contrast, single vital rates were affected by mixed effects of all three factors studied. In particular, genetic diversity and population size mainly affected growth related transitions, while habitat quality mainly affected germination, stasis and mortality. Hence, we conclude that different factors affect different parts of the life cycle. However, not all of the effects on single vital rates will translate into effects on the overall population growth rate, and judging their effects on long-term population performance is not possible without following the whole life cycle of the species.

Acknowledgements We would like to thank to the workers in the protected areas of the Czech and Slovak Republic where L. sibirica occurs for providing us with information about its localities. We obtained permission from the Ministries of the Environment of the Czech and Slovak Republic to enter the localities of populations of this species, and for its manipulation. This study was supported by grant MŠMT 2B06178 and also partly by grants MŠMT 0021620828 and AV0Z60050516.

References Ågren, J. (1996): Population size, pollinator limitation, and seed set in the self-incompatible herb Lythrum salicaria. Ecol., 77: 1779-1790. Ågren, J., Ehrlén, J. & Solbreck, C. (2008): Spatio-temporal variation in fruit production and seed predation in a perennial herb influenced by habitat quality and population size. J. Ecol., 96: 334-345. Alvarez-Buylla, E.R. & Slatkin, M. (1994): Finding confidence limits on population growth rates: three real examples revised. Ecol., 75: 255-260. Barrett, S.C. & Kohn, J.R. (1991): Genetic and evolutionary consequences of small population size in plants: implications for conservation. Genetics and Conservation of Rare Plants (eds D.A. Falk & K.E. Holsinger), pp. 3-30. Oxford University Press, New York. Brys, R., Jaquemyn, H., Endels, P., van Rossum, F., Hermy, M., Triest, L., de Bruyn, L. & de Blust, G. (2004): Reduced reproductive success in small populations of the self-incompatible Primula vulgaris. J. Ecol., 92: 5-14. Brys, R., Jaquemyn, H., Endels, P., de Blust, G. & Hermy, M. (2005): Effect of habitat deterioration on population dynamics and extinction risks in a previously common perennial. Conserv. Biol., 19: 1633-1643. Cariveau, D., Irwin, R.E., Brody, A.K., Garcia-Mayeya, L.S. & von der Ohe, M. (2004): Direct and indirect effects of pollinators and seed predators to selection on plant and floral traits. Oikos, 104: 15-26. Caswell, H. (2001): Matrix population models, construction, analysis, and interpretation. Sinauer Associates, Sunderland, MA. Charlesworth, D. & Charlesworth, B. (1987): Inbreeding depression and its evolutionary consequences. Annu. Rev. Ecol. Syst., 18: 237-268. Cochran, M.E. & Ellner, S. (1992): Simple Methods for Calculating Age-Based Life History Parameters for Stage-Structured Populations. Ecol. Monogr., 62: 345-364.

117 Chapter IV

Colling, G., Matthies, D. & Reckinger, C. (2002): Population structure and establishment of the threatened long-lived perennial Scorzonera humilis in relation to environment. J. Appl. Ecol., 39: 310-320. Colling, G. & Matthies, D. (2006): Effects of habitat deterioration on population dynamics and extinction risk of an endangered, long-lived perennial herb (Scorzonera humilis). J. Ecol., 94: 959-972. Council of European Communities (1992): Council Directive 92/43/EEC of 21 May on the conservation of natural habitats and of wild fauna and flora. Off. J. Eur. Communities, 35: 7- 50. Dahlgren, J.P. & Ehrlén, J. (2009): Linking environmental variation to population dynamic of a forest herb. J. Ecol., 97: 666-674. De Kroon, H., van Groenendael, J. & Ehrlén, J. (2000): Elasticities: a review of methods and model limitations. Ecol., 81: 607-618. De Vere, N., Jongejans, E., Plowman, A. & Williams, E. (2009): Population size and habitat quality affect genetic diversity and fitness in the clonal herb Cirsium dissectum. Oecologia, 159: 59-68. Diekmann, M. (2003): Species indicator values as an important tool in applied plant ecology – a review. Basic Appl. Ecol., 4: 493-506. Doust, L.L. (1981): Population dynamics and local specialization in a clonal perennial (Ranunculus repens): I. The dynamics of ramets in contrasting habitats. J. Ecol., 69: 743-755. Dudash, M.R. & Fenster, C.B. (2000): Inbreeding and outbreeding depression in fragmented populations. Genetics, Demography and Viability of Fragmented Populations (eds A.G. Young & G.M. Clarke), pp. 35-53. Cambridge University Press, Cambridge. Efron, B. & Tibshirani, R.J. (1994): Introduction to the bootstrap. Chapman and Hall, New York. Ehrlén, J. & Lehtilä, K. (2002): How perennial are perennial plants? Oikos, 98: 308-322. Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W. & Paulißen, D. (1992): Zeigerwerte von Pflanzen in Mitteleuropa. 2nd ed. Scripta Geobotanika, 18: 1-258. Ellenberg H. & Leuschner C. (2010): Vegetation Mitteleuropas mit den Alpen. UTB, Ulmer. Ellstrand, N.C. & Elam, D.C. (1993): Population genetic consequences of small population size: implication for plant conservation. Annu. Rev. Ecol. Syst., 24: 217-242. Feráková, V., Maglocký, Š. & Marhold, K. (2001): Červený zoznam papraďorastov a semenných rastlín Slovenska (december 2001). In Baláž, D., Marhold, K. & Urban, P. (eds) Červený zoznam rastlín a živočíchov Slovenska. Ochr. Prír., Banská Bystrica, 20: 48–81. Fischer, M. & Matthies, D. (1998): Effects of population size on performance in the rare plant Gentianella germanica. J. Ecol., 86: 195-204. Hanski, I. & Ovaskainen, O. (2002): Extinction debt at extinction threshold. Conserv. Biol., 16: 666-673. Hegi, G. (1929): Illustrierte Flora von Mittel – Europa, Ed. 1. Vol. 6/2. München, pp. 796-799. Heinken-Šmídová, A. & Münzbergová, Z. (in press): Population dynamics of the endangered, long-lived perennial species, Ligularia sibirica. Folia Geobot. Hendrych, R. (2003): Poznatky o druhu Ligularia sibirica v Čechách. Preslia, 75: 39-69. Heywood, V.H. & Iriondo, J.M. (2003): Plant conservation: old problems, new perspectives. Biol. Conserv., 113: 321-335. Jennersten, O. (1988): Pollination in Dianthus deltoides (Caryophyllaceae) Effects of habitat fragmentation on visitation and seed set. Conserv. Biol., 2: 359-366. Käfer, J. & Witte, M.J-P. (2004): Cover-weighted averaging of indicator values in vegetation analyses. J. Veg. Sci., 15: 647-652. Kéry, M., Matthies, D. & Fischer, M. (2001): The effect of plant population size on the interactions between the rare plant Gentiana cruciata and its specialized herbivore Maculinea rebeli. J. Ecol., 89: 418-427. Kolb, A., Dahlgren, J. & Ehrlén, J. (2010): Population size affects vital rates but not population growth rate of a perennial plant. Ecol., 91: 3210-3217. Kukk, Ü. (2003) The distribution of Ligularia sibirica (L.) Cass. in Estonia and changes in its population. Biul. Ogrodów Bot., 12: 11-22.

118 Chapter IV

Leimu, R., Mutikainen, P., Koricheva, J. & Fischer, M. (2006): How general are positive relationships between plant population size, fitness and genetic variation? J. Ecol., 94: 942- 952. Liu, J-Q. (2004): Uniformity of karyotypes in Ligularia (Asteraceae: Senecioneae), a highly diversified genus of the eastern Qinghai–Tibet Plateau highlands and adjacent areas. Bot. J. Linn. Soc., 144: 329-342. Luijten, S.H., Dierick, A., Oostermeijer, J.G.B., Raijmann, L.E.L. & den Nijs, J.C.M. (2000): Population size, genetic variation and reproductive success in a rapidly declining, self- incompatible perennial (Arnica montana) in The Netherlands. Conserv. Biol., 14: 1776-1786. Matlab (2006): Matlab version 7.3.0.267. Mathworks Inc., Natick, MA, USA. Moravec, J. (1994): Fytocenologie, Academia, Praha. (In Czech) Münzbergová, Z. (2005): Determinants of species rarity: population growth rates of species sharing the same habitat. Am. J. Bot., 92: 1987-1994. Münzbergová, Z. (2006a): Effect of population size on the prospect of species survival. Folia Geobot. 41: 137-150. Münzbergová, Z. (2006b): Ploidy level interacts with population size and habitat conditions to determine degree of herbivory damage in plant populations. Oikos, 115: 443-452. Münzbergová, Z. (2007): Population dynamics of diploid and hexaploid populations of a perennial herb. Ann. Bot., 100: 1259-1270. Oostermeijer, J.G.B., van Eijck, M.W., van Leeuwen, N.C. & den Nijs, J.C.M. (1995): Analysis of the relationship between allozme heterozygosity and fitness in the rare Gentiana pneumonanthe L. J. Evol. Biol., 8: 739-759. Oostermeijer, J.G.B., Brugman, M.L., de Boer, E.R. & den Nijs, H.C.M. (1996): Temporal and spatial variation in the demography of Gentiana pneumonanthe, a rare perennial herb. J. Ecol., 84: 153-166. Oostermeijer, J.G.B., Luijten, S.H., Křenová, Z.V. & den Nijs, J.C.M. (1998): Relationship between population and habitat characteristics and reproduction of the rare Gentiana pneumonanthe L. Conserv. Biol., 12: 1042-1053. Oostermeijer, J.G.B., Luijten, S.H. & den Nijs, J.C.M. (2003): Integrating demographic and genetic approaches in plant conservation. Biol. Conserv., 113: 389-398. Östergård, H. & Ehrlén, J. (2005): Among population variation in specialist and generalist seed predation – the importance of host plant distribution alternative hosts and environmental variation. Oikos, 111: 39-46. Ouborg, N.J. & van Treuren, R. (1995): Variation in fitness-related characters among small and large populations of Salvia pratensis. J. Ecol., 83: 369-380. Ouborg, N.J., Vergeer, P. & Mix, C. (2006): The rough edges of the conservation genetics paradigm for plants. J. Ecol., 94: 1233-1248. Paschke, M., Abs, C. & Schmid, B. (2002): Relationship between population size, allozyme variation, and plant performance in the narrow endemic Cochleria bavarica. Conserv. Genet., 3: 131-144. Primack, R.B. (2002): Essentials of conservation biology. Ed. 3. Sinauer Associates, Suderland. Procházka, F. (ed) (2001): Černý a červený seznam cévnatých rostlin České republiky (stav v roce 2000). Příroda, Praha 18: 1–166. Schaffers, A.P. & Sýkora, K.V. (2000): Reliability of Ellenberg indicator values for moisture, nitrogen and soil reaction: a comparison with field measurements. J. Veg. Sci., 11: 225-244. Schemske, D.W., Husband, B.C., Ruckelhaus, M.H., Goodwille, C., Parker, I.M. & Bishop, J.G. (1994): Evaluating approaches to the conservation of rare and endangered plants. Ecol., 75: 584-606. Schleuning, M. & Matthies, D. (2009): Habitat change and plant demography: Assessing the extinction risk of a formerly common grassland perennial. Conserv. Biol., 23: 174-183. Silvertown, J. & Franco, M. (1993): Plant demography and habitat: a comparative approach. Plant Species Biol., 8: 67-73. Sletvold, N. & Rydgren, K. (2007): Population dynamics in Digitalis purpurea: the interaction of disturbance and seed bank dynamics. J. Ecol., 95: 1346-1359. Slavík, B. (2004): Ligularia Cass. In Slavík, B. & Štěpánková, J. (eds) Květena České republiky, Vol. 7. Academia, Praha, pp 306–309.

119 Chapter IV

S-Plus (2000): Professional edition for Windows, release 2. (1999). MathSoft, Inc., Massachusetts Šmídová, A. (2006): Popelivka sibiřská (Ligularia sibirica) na Jestřebských slatinách. Severočeskou Přír., 38: 11–17. Šmídová, A. (2009): Studie faktorů ovlivňující dynamiku druhu Ligularia sibirica na Slovensku. Prír Nízkych Tatier, Banská Bystrica, 2: 55–60. Šmídová, A., Münzbergová, Z. & Plačková, I. (2011): Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae). Flora, 206: 151-157. Ter Braak, C.J.F. & Šmilauer, P. (2002): CANOCO Reference manual and CanoDraw for Windows User’s Guide: Software for canonical community ordination, Version 4.5. Microcomputer Power, Ithaca. Vergeer, P.V., Rengelink, R., Copal, A. & Ouborg, N.J. (2003): The interacting effects of genetic variation, habitat quality and population size in performance of Succisa pratensis. J. Ecol., 91: 18-26. Young, A.G., Boyle, T. & Brown, T. (1996): The population genetic consequences of habitat fragmentation for plants. Trends Ecol. Evol., 11: 413-418.

120 Chapter IV

Electronic Supplementary Material

Table S1: Relationship between the scores of vegetation relevés from the localities on the first DCA axis and the Ellenberg indicator values. Significant values are in bold, P < 0.05. Strongly correlated values are in italics

Nutrient Moisture Reaction Light availability Moisture -0.56

Reaction 0.46 -0.32

Light -0.79 0.41 -0.1 Vegetation -0.94 0.48 -0.34 0.86 composition

121 Chapter IV

s. Significance values are values0.05s. Significance in bold, P < Table S2: The relationship between dependend variable between dependend relationship The Table S2:

122 Chapter IV

a57 a16 a36 a56 a67 1 2 3 4 5 6 7

a21 a32 a43 a54 a65 a76

a44 a55 a66 a77

a a a 31 53 75

Fig. S1: Life-cycle diagram of the seven stages of Ligularia sibirica: (1) seed bank first year, (2) seed bank second year, (3) seedlings, (4) juveniles, (5) vegetative adults, (6) fertile adults and (7) dormant adult individuals. (Heinken-Šmídová A. & Münzbergová Z., in press)

1.8

1.6

1.4

λ 1.2

1

0.8

/1 /2 /3 /4 /5 /1 /2 /3 /1 R /1 N N N N N C C C S S P Z Z Z Z Z Z Z Z Z K K C C C C C C C C C S S

Fig. S2: Stochastic population growth rate of L. sibirica populations with 95% confidence intervals. The horizontal line indicate population growth rate (λ) = 1, i.e. stable population. See Table 1 for abbreviations of the study localities (Heinken-Šmídová A. & Münzbergová Z., in press)

123 Chapter IV

Fig. S3: DCA ordination plot of the vegetation samples. The first axis represents a gradient from expansive species to species of calcareous fens and was used to describe habitat conditions in the analyses. The first horizontal axis explains 13.1% of the total variation in the dataset. The second vertical axis explains 6.8% of the total variation in the dataset. The species with the greatest weights were selected for display

Species list: Aln glu – Alnus glutinosa, Ang syl = Angelica sylvestris, Bra mil = Brachythecium mildeanum, Bra riv = Brachythecium rivulare, Bry pse = Bryum pseudotriquetrum, Cal cus = Calliergonella cuspidata, Cal epi = Calamagrostis epigejos, Cam ste = Campyllium stellatum, Car acu = Carex acutiformis, Car dav = Carex davalliana, Car pan1 = Carex panicea, Car pan2 = Carex paniculata, Cir ole = Cirsium oleraceum, Cir pal = Cirsium palustre, Cir riv = Cirsium rivulare, Cli den = Climacium dendroides, Cre pal = Crepis paludosa, Des ces = Deschampsia cespitosa, Equ pal = Equisetum palustre, Eup can = Eupatorium cannabinum, Fes rub = Festuca rubra, Fil ulm = Filipendula ulmaria, Fis adi = Fissidens adianthoides, Fra aln = Frangula alnus, Gal uli = Galium uliginosum, Geu riv = Geum rivale, Lat pra = Lathyrus pratensis, Lig sib = Ligularia sibirica, Lys vul = Lysimachia vulgaris, Mol cae = Molinia caerulea, Par pal = Parnassia palustris, Phr aus = Phragmites australis, Pla ela = Plagiomnium elatum, Pot ere = Potentilla erecta, Ran acr = Ranunculus acris, San off = Sanguisorba officanalis, Sel car = Selinum carvifolia, Suc pra = Succisa pratensis, Val dio = Valeriana dioica

124 Chapter IV

(A)

(B)

Fig. S4:Relationship between the scores of vegetation relevés from the localities on the first DCA axis and the Ellenberg indicator values of nutrient availability (Fig. A; r = -0.94; P < 0.001) and light (Fig. B; r = 0.86; P < 0.001). See Table 1 for abbreviations of the study localities

125

CONCLUSIONS

Conclusions

Conclusions

In this thesis I aimed to investigate how the possible main factors and their interactions affect population viability of the rare relict plant species Ligularia sibirica (L.) Cass..

The main results of my study indicate that most of the investigated populations of L. sibirica, both small and large, are not declining and that they have very low extinction probabilities. The results also indicate that all populations have low adult mortality (i.e. a long life-span of adult individuals in the range 13-135 years). However, there is one population which is currently declining and two populations showed a tendency towards decline. All those declining populations are growing in nutrient- and especially nitrogen-rich habitats in northern Bohemia in the Czech Republic and have a high extinction risk. Those populations show features of remnant populations (i.e. the persistence of populations in severe conditions in spite of no reproduction) (Chapter I). In another part of my study I found that populations of L. sibirica still preserved high levels of genetic variability. Genetic variability is mainly within populations, while the level of genetic differentiation between populations is low. The results also revealed a significant positive correlation between genetic diversity and population size. Exceptions are small populations in northern Bohemia, the Czech Republic, which are of lower genetic variability. Their lower genetic variability is probably caused by destruction of habitats by melioration during the last century and subsequent decrease of population size of L. sibirica. Patterns of genetic diversity, in the meaning of high level of genetic variability and low genetic differentiation, suggest that the recent distribution is a result of stepwise postglacial migration of the species and subsequent natural fragmentation (Chapter II). By the description of habitat requirements of L. sibirica I found out that L. sibirica has the ability to grow and persist in a wide range of plant communities with preference to the Caricion davallianae alliance or transitions between Caricion davallianae and Molinion caeruleae communities. These sites are characterized by a high level of ground water, neutral to alkaline pH, high content of calcium carbonate, phosphorus deficiency and a low biomass production. Further, the results showed that L. sibirica has the ability to grow and compete under various nutrient regimes. The findings also indicate that adult individuals of L. sibirica which grow in those communities are of small sizes (i.e. with low numbers of leaf rosettes, flower heads and low seed production), while individuals that persist in degraded or abandoned habitats are of larger size (Chapter III). Those results together with the results of study on population dynamics of L. sibirica (Chapter I) indicate that viable populations of the species seem to be independent of the size of individuals (Chapter III).

128 Conclusions

In the last chapter (Chapter IV) I used the findings of previous chapters (Chapter I, II and III) and one additional factor, population size, to determine which of the studied factors or their combination have important influence on population viability of L. sibirica. The results indicate that the population growth rate of L. sibirica and the importance of single transitions to population growth rate (elasticity) are mainly affected by habitat conditions. Single vital rates, however, were affected by all three studied factors. In particular, genetic diversity and population size mainly affected growth related transitions, whereas habitat conditions mainly affected germination, stasis and mortality.

Overall, the results of this thesis showed that different factors and/or their combination may affect different parts of the life cycle and different single vital rates, and that not all of these effects will be translated into effects on the overall population growth rate. The findings of my study thus emphasize the importance of complex studies of species biology. In particular, detailed demographic studies are necessary to assess population viability. They may serve as an early warning system for species long before an obvious decline occurs in the populations and thus help to give advices for long-term protection of the species. Findings showed that habitat quality (i.e. high water table and low nutrient availability as to be found in intact fens) is the most important predictor of population viability. I conclude that population size and genetic diversity in this plant species which is growing in fragmented populations presumably already since the early postglacial are not yet a major problem for the conservation of L. sibirica. Conservation activities should thus focus on preserving and – where necessary – restoring habitat conditions of the localities.

The presented study raises some questions for future research on L. sibirica. These are based on the fact that L. sibirica is belonging to the species that are naturally rare in a specific area, and occurs in small, isolated populations already for a long time. Firstly, an interesting question is why viable populations of L. sibirica grow in such a wide variety of habitats and yet many apparently suitable habitats are not occupied. This question also arises for many other rare plant species. Secondly, the role of management for conservation of L. sibirica was not studied experimentally in my thesis. Thus advices for concrete management actions as intensity of mowing on the different sites are difficult to give. To ensure this, the establishment of different management variants on every site would be necessary.

129

Curriculum vitae

Curriculum vitae

Personal data: Name: Anna Heinken-Šmídová (nee Šmídová) Born: 13. 11. 1977, Česká Lípa, Czech Republic

Education: Since 2004 Palacký University, Faculty of Science, Department of Botany, Olomouc, Czech Republic; Ph.D. study in botany, thesis: The study of factors influencing dynamics of an endangered plant species – Ligularia sibirica (supervisor: Doc. RNDr. Zuzana Münzbergová Ph.D.) 2001 – 2004 parallel pedagogical training in Environmental Education 1998 – 2004 Palacký University, Faculty of Science, Department of Ecology and Environmental Sciences, Olomouc, Czech Republic, Bachelor thesis: „The natural characteristic and the management plan of prepared specially protected area Meanders of the river Ploučnice“ (supervisor: Mgr. Vlastimil Rybka Ph.D.). Master thesis: “Ecobiology of Ligularia sibirica (supervisor: Mgr. Vlastimil Rybka Ph.D.)

Work experience: Since 2011 Maternity leave 2006 – 2011 Academy of Science of the Czech Republic, Institute of Botany, Department of Population Ecology, research assistant 2006 – 2011 Participation on the project: Priorities of conservation of vascular plant species. Grant project supported by Ministry of Education of The Czech Republic 2B06178. (member of research team) 2004 – 2005 Participation on the project: Analysis of biodiversity in the Landscape Protected Area Bílé Karpaty as a background for defining of new zonation and appropriate management of valuable areas (program Biosphere – SE/theme VaV/620/12/03) Since 2002 Monitoring and participation on the proposal of the methodology for monitoring of an important plant species of the EU Habitats Directive (Annex II) - Ligularia sibirica 2002 – 2003 Participation on projects: Coexistence of species in species rich meadows (206/97/0937 Grant Agency of the Czech Republic) and the mechanisms of the long-term dynamics in species rich meadows working on a small-scale in the Landscape Protected Area Bílé Karpaty 2001 – 2004 Mapping of habitats in the system of the “NATURA 2000” - network (Liberec region) 131 Curriculum vitae

Stays abroad:

2006 three months at the Institute of Biochemistry and Biology, University of Potsdam, Germany, working with Prof. Dr. Markus Fischer and Roosa Leimu, Ph.D., scholarship from the Student Mobility Support Programme.

Publications in international peer-reviewed journals: Heinken-Šmídová A. & Münzbergová Z. (2012): Population dynamics of the endangered, long- lived perennial species, Ligularia sibirica (in press Folia Geobotanica) Šmídová A., Münzbergová Z. & Plačková I. (2011): Genetic diversity of a relict plant species, Ligularia sibirica (L.) Cass. (Asteraceae). Flora, 206: 151-157.

Other publications: Heinken-Šmídová A. & Münzbergová Z. Faktory ovlivňující populační dynamiku kriticky ohroženého druhu Ligularia sibirica (in press Příroda) Šmídová A. (2009): Studie faktorů ovlivňující dynamiku duhu Ligularia sibirica na Slovensku. Priroda Nizkych Tatier, Banská Bystrica, 2: 55-60. Šmídová A. (2006): Popelivka sibiřská (Ligularia sibirica) na Jestřebských slatinách. Severočes. Přír., Litoměřice, 38: 11-17.

Participation in conferences: Heinken-Šmídová A. & Münzbergová Z. (2010): Genetická diversita populací kriticky ohroženého druhu, Ligularia sibirica. — Konference ČBS: “Evoluční aspekty biologie rostlin.” Praha, 27. – 28. 11. 2010 (Oral presentation in Czech) Šmídová A. & Münzbergová Z. (2010): Habitat conditions are better determinants of population performance of perennial species Ligularia sibirica, than its size. — Konference 23rd annual meeting of the GfÖ section Plant Population Biology: “Plant population biology crossing borders.” Nijmegen, Netherlands, 13. – 15. 5. 2010 (Oral presentation in English) Šmídová A. & Münzbergová Z. (2009): Factors influencing population dynamics of a critically endangered plant species, Ligularia sibirica. — Konference “2nd European Congress of Conservation Biology.” Prague, Czech Republic, 1. – 5. 9. 2009 (Oral presentation in English) Šmídová A. (2008): Studie faktorů ovlivňující dynamiku kriticky ohroženého rostlinného druhu Ligularia sibirica na Slovensku. — Konference “Priroda Nizkych Tatier.” Liptovský Ján, Slovensko, 24. – 25. 9. 2008 (Oral presentation in Czech)

132 Curriculum vitae

Šmídová A. & Münzbergová Z. (2008): Population genetic diversity of an endangered plant species, Ligularia sibirica. — Konference “21st annual meeting of the GfÖ section Plant Population Biology: “Plant Population Biology for the coming decade.” Luxembourg, Luxembourg, 1. – 3. 5. 2008 (Poster presentation) Šmídová A. & Münzbergová Z. (2008): Genetická diversita populací kriticky ohroženého druhu, Ligularia sibirica. — Konference ČSPE: “Ekologie v 21 století.” Třeboň, 25. – 27. 4. 2008 (Poster presentation) Šmídová A. & Münzbergová Z. (2006): Studie faktorů ovlivňující dynamiku druhu Ligularia sibirica. — Konference ČBS: “Vodní a mokřadní rostliny – taxony, společenstva, vztahy.” Praha, 24 – 25. 11. 2006 (Oral presentation in Czech) Šmídová A. & Münzbergová Z. (2006): Faktory ovlivňující dynamiku populací popelivky sibiřské (Ligularia sibirica) na Jestřebských slatinách. — Konference: “Ohrožené a jinak významné rostliny a biotopy severních Čech.” Ústí nad Labem, 16 – 17. 9. 2006 (Oral presentation in Czech) Šmídová A. & Münzbergová Z. (2006): Study of factors influencing population dynamics of an endangered plant species, Ligularia sibirica. — Konference “19th annual meeting of the GfÖ section Plant Population Biology: “Biotic interactions.” Halle/Saale, Germany, 24 – 27. 5. 2006 (Poster presentation)

133